init

experiments/baseline_lm_lc.py

@@ -1,6 +1,8 @@
 import os
 import json
+import gc
 
+import torch
 import pandas as pd
 from datasets import load_dataset
 from lmppl import EncoderDecoderLM, LM, OpenAI
@@ -17,29 +19,29 @@ prompt_dict = {
 data = load_dataset("cardiffnlp/relentless", split="test")
 full_result = []
 for lm, ppl_class, batch, pretty_name in [
-    ("t5-small", EncoderDecoderLM, 256, "T5\textsubscript{SMALL}"),
-    ("t5-base", EncoderDecoderLM, 128, "T5\textsubscript{BASE}"),
-    ("t5-large", EncoderDecoderLM, 32, "T5\textsubscript{LARGE}"),
-    ("t5-3b", EncoderDecoderLM, 1, "T5\textsubscript{XL}"),
-    ("t5-11b", EncoderDecoderLM, 1, "T5\textsubscript{XXL}"),
-    ("google/flan-t5-small", EncoderDecoderLM, 256, "Flan-T5\textsubscript{SMALL}"),
-    ("google/flan-t5-base", EncoderDecoderLM, 128, "Flan-T5\textsubscript{BASE}"),
-    ("google/flan-t5-large", EncoderDecoderLM, 32, "Flan-T5\textsubscript{LARGE}"),
-    ("google/flan-t5-xl", EncoderDecoderLM, 1, "Flan-T5\textsubscript{XL}"),
-    ("google/flan-t5-xxl", EncoderDecoderLM, 1, "Flan-T5\textsubscript{XXL}"),
     ("google/flan-ul2", EncoderDecoderLM, 1, "Flan-UL2"),
-    ("facebook/opt-125m", LM, 256, "OPT\textsubscript{125M}"),
-    ("facebook/opt-350m", LM, 128, "OPT\textsubscript{350M}"),
-    ("facebook/opt-1.3b", LM, 1, "OPT\textsubscript{1.3B}"),
-    ("facebook/opt-2.7b", LM, 1, "OPT\textsubscript{2.7B}"),
-    ("facebook/opt-6.7b", LM, 1, "OPT\textsubscript{6.7B}"),
-    ("facebook/opt-13b", LM, 1, "OPT\textsubscript{13B}"),
+    ("google/flan-t5-xxl", EncoderDecoderLM, 1, "Flan-T5\textsubscript{XXL}"),
+    ("google/flan-t5-xl", EncoderDecoderLM, 1, "Flan-T5\textsubscript{XL}"),
+    ("google/flan-t5-large", EncoderDecoderLM, 32, "Flan-T5\textsubscript{LARGE}"),
+    ("google/flan-t5-base", EncoderDecoderLM, 128, "Flan-T5\textsubscript{BASE}"),
+    ("google/flan-t5-small", EncoderDecoderLM, 256, "Flan-T5\textsubscript{SMALL}"),
+    ("t5-11b", EncoderDecoderLM, 1, "T5\textsubscript{XXL}"),
+    ("t5-3b", EncoderDecoderLM, 1, "T5\textsubscript{XL}"),
+    ("t5-large", EncoderDecoderLM, 32, "T5\textsubscript{LARGE}"),
+    ("t5-base", EncoderDecoderLM, 128, "T5\textsubscript{BASE}"),
+    ("t5-small", EncoderDecoderLM, 256, "T5\textsubscript{SMALL}"),
+    # ("facebook/opt-66b", LM, 1, "OPT\textsubscript{66B}"),
     ("facebook/opt-30b", LM, 1, "OPT\textsubscript{30B}"),
-    ("facebook/opt-66b", LM, 1, "OPT\textsubscript{66B}"),
-    ("facebook/opt-iml-1.3b", LM, 1, "OPT-IML\textsubscript{1.3B}"),
+    ("facebook/opt-13b", LM, 1, "OPT\textsubscript{13B}"),
+    ("facebook/opt-6.7b", LM, 1, "OPT\textsubscript{6.7B}"),
+    ("facebook/opt-2.7b", LM, 1, "OPT\textsubscript{2.7B}"),
+    ("facebook/opt-1.3b", LM, 1, "OPT\textsubscript{1.3B}"),
+    ("facebook/opt-350m", LM, 128, "OPT\textsubscript{350M}"),
+    ("facebook/opt-125m", LM, 256, "OPT\textsubscript{125M}"),
     ("facebook/opt-iml-30b", LM, 1, "OPT-IML\textsubscript{30B}"),
-    ("facebook/opt-iml-max-1.3b", LM, 1, "OPT-IML\textsubscript{MAX-1.3B}"),
+    ("facebook/opt-iml-1.3b", LM, 1, "OPT-IML\textsubscript{1.3B}"),
     ("facebook/opt-iml-max-30b", LM, 1, "OPT-IML\textsubscript{MAX-30B}"),
+    ("facebook/opt-iml-max-1.3b", LM, 1, "OPT-IML\textsubscript{MAX-1.3B}"),
     # ("davinci", OpenAI, None, "GPT-3\textsubscript{davinci}")
 ]:
     os.makedirs(f"experiments/results/lm_lc/{os.path.basename(lm)}", exist_ok=True)
@@ -82,6 +84,9 @@ for lm, ppl_class, batch, pretty_name in [
         tmp = pd.DataFrame([true_rank, prediction], index=['true', 'pred']).T
         cor = tmp.corr("spearman").values[0, 1]
         full_result.append({"model": pretty_name, "relation_type": d['relation_type'], "correlation": cor})
+    del scorer
+    gc.collect()
+    torch.cuda.empty_cache()
 
 df = pd.DataFrame(full_result)
 models = df['model'].unique()

experiments/baseline_lm_qa.py

@@ -1,6 +1,8 @@
 import os
 import json
+import gc
 
+import torch
 import pandas as pd
 from datasets import load_dataset
 from lmppl import EncoderDecoderLM, LM, OpenAI
@@ -17,30 +19,30 @@ prompt_dict = {
 data = load_dataset("cardiffnlp/relentless", split="test")
 full_result = []
 for lm, ppl_class, batch, pretty_name in [
-    ("t5-small", EncoderDecoderLM, 256, "T5\textsubscript{SMALL}"),
-    ("t5-base", EncoderDecoderLM, 128, "T5\textsubscript{BASE}"),
-    ("t5-large", EncoderDecoderLM, 32, "T5\textsubscript{LARGE}"),
-    ("t5-3b", EncoderDecoderLM, 1, "T5\textsubscript{XL}"),
-    ("t5-11b", EncoderDecoderLM, 1, "T5\textsubscript{XXL}"),
-    ("google/flan-t5-small", EncoderDecoderLM, 256, "Flan-T5\textsubscript{SMALL}"),
-    ("google/flan-t5-base", EncoderDecoderLM, 128, "Flan-T5\textsubscript{BASE}"),
-    ("google/flan-t5-large", EncoderDecoderLM, 32, "Flan-T5\textsubscript{LARGE}"),
-    ("google/flan-t5-xl", EncoderDecoderLM, 1, "Flan-T5\textsubscript{XL}"),
-    ("google/flan-t5-xxl", EncoderDecoderLM, 1, "Flan-T5\textsubscript{XXL}"),
     ("google/flan-ul2", EncoderDecoderLM, 1, "Flan-UL2"),
-    ("facebook/opt-125m", LM, 256, "OPT\textsubscript{125M}"),
-    ("facebook/opt-350m", LM, 128, "OPT\textsubscript{350M}"),
-    ("facebook/opt-1.3b", LM, 1, "OPT\textsubscript{1.3B}"),
-    ("facebook/opt-2.7b", LM, 1, "OPT\textsubscript{2.7B}"),
-    ("facebook/opt-6.7b", LM, 1, "OPT\textsubscript{6.7B}"),
-    ("facebook/opt-13b", LM, 1, "OPT\textsubscript{13B}"),
+    ("google/flan-t5-xxl", EncoderDecoderLM, 1, "Flan-T5\textsubscript{XXL}"),
+    ("google/flan-t5-xl", EncoderDecoderLM, 1, "Flan-T5\textsubscript{XL}"),
+    ("google/flan-t5-large", EncoderDecoderLM, 32, "Flan-T5\textsubscript{LARGE}"),
+    ("google/flan-t5-base", EncoderDecoderLM, 128, "Flan-T5\textsubscript{BASE}"),
+    ("google/flan-t5-small", EncoderDecoderLM, 256, "Flan-T5\textsubscript{SMALL}"),
+    ("t5-11b", EncoderDecoderLM, 1, "T5\textsubscript{XXL}"),
+    ("t5-3b", EncoderDecoderLM, 1, "T5\textsubscript{XL}"),
+    ("t5-large", EncoderDecoderLM, 32, "T5\textsubscript{LARGE}"),
+    ("t5-base", EncoderDecoderLM, 128, "T5\textsubscript{BASE}"),
+    ("t5-small", EncoderDecoderLM, 256, "T5\textsubscript{SMALL}"),
+    # ("facebook/opt-66b", LM, 1, "OPT\textsubscript{66B}"),
     ("facebook/opt-30b", LM, 1, "OPT\textsubscript{30B}"),
-    ("facebook/opt-66b", LM, 1, "OPT\textsubscript{66B}"),
-    ("facebook/opt-iml-1.3b", LM, 1, "OPT-IML\textsubscript{1.3B}"),
+    ("facebook/opt-13b", LM, 1, "OPT\textsubscript{13B}"),
+    ("facebook/opt-6.7b", LM, 1, "OPT\textsubscript{6.7B}"),
+    ("facebook/opt-2.7b", LM, 1, "OPT\textsubscript{2.7B}"),
+    ("facebook/opt-1.3b", LM, 1, "OPT\textsubscript{1.3B}"),
+    ("facebook/opt-350m", LM, 128, "OPT\textsubscript{350M}"),
+    ("facebook/opt-125m", LM, 256, "OPT\textsubscript{125M}"),
     ("facebook/opt-iml-30b", LM, 1, "OPT-IML\textsubscript{30B}"),
-    ("facebook/opt-iml-max-1.3b", LM, 1, "OPT-IML\textsubscript{MAX-1.3B}"),
+    ("facebook/opt-iml-1.3b", LM, 1, "OPT-IML\textsubscript{1.3B}"),
     ("facebook/opt-iml-max-30b", LM, 1, "OPT-IML\textsubscript{MAX-30B}"),
-    ("davinci", OpenAI, None, "GPT-3\textsubscript{davinci}")
+    ("facebook/opt-iml-max-1.3b", LM, 1, "OPT-IML\textsubscript{MAX-1.3B}"),
+    # ("davinci", OpenAI, None, "GPT-3\textsubscript{davinci}")
 ]:
     os.makedirs(f"experiments/results/lm_qa/{os.path.basename(lm)}", exist_ok=True)
     scorer = None
@@ -80,6 +82,9 @@ for lm, ppl_class, batch, pretty_name in [
         tmp = pd.DataFrame([true_rank, prediction], index=['true', 'pred']).T
         cor = tmp.corr("spearman").values[0, 1]
         full_result.append({"model": pretty_name, "relation_type": d['relation_type'], "correlation": cor})
+    del scorer
+    gc.collect()
+    torch.cuda.empty_cache()
 
 df = pd.DataFrame(full_result)
 models = df['model'].unique()