From 1868e86e2ac1b5fcf72106187ef8936ec5803070 Mon Sep 17 00:00:00 2001
From: Juhan Bae <pomonam15@gmail.com>
Date: Tue, 9 Jul 2024 14:17:43 -0400
Subject: [PATCH] Modify default batch_size

---
 examples/imagenet/README.md  | 1 -
 examples/imagenet/analyze.py | 8 +-------
 2 files changed, 1 insertion(+), 8 deletions(-)

diff --git a/examples/imagenet/README.md b/examples/imagenet/README.md
index 6726862..e30ca71 100644
--- a/examples/imagenet/README.md
+++ b/examples/imagenet/README.md
@@ -16,7 +16,6 @@ To compute pairwise influence scores on 1000 query data points using the `ekfac`
 ```bash
 python analyze.py --dataset_dir PATH_TO_IMAGENET \
     --query_gradient_rank -1 \
-    --factor_batch_size 512 \
     --query_batch_size 100 \
     --train_batch_size 256 \
     --factor_strategy ekfac
diff --git a/examples/imagenet/analyze.py b/examples/imagenet/analyze.py
index aa2517c..255a65e 100644
--- a/examples/imagenet/analyze.py
+++ b/examples/imagenet/analyze.py
@@ -34,12 +34,6 @@ def parse_args():
         default=-1,
         help="Rank for the low-rank query gradient approximation.",
     )
-    parser.add_argument(
-        "--factor_batch_size",
-        type=int,
-        default=512,
-        help="Batch size for computing factors.",
-    )
     parser.add_argument(
         "--query_batch_size",
         type=int,
@@ -104,7 +98,7 @@ def main():
     analyzer.fit_all_factors(
         factors_name=factors_name,
         dataset=train_dataset,
-        per_device_batch_size=args.factor_batch_size,
+        per_device_batch_size=None,
         factor_args=factor_args,
         overwrite_output_dir=False,
     )