From e710cc041f8147c8894c86366c33bff1407ffe01 Mon Sep 17 00:00:00 2001
From: Florian <26519986+F5D7@users.noreply.github.com>
Date: Fri, 19 Aug 2022 14:33:36 +0200
Subject: [PATCH] added training script note

The behavior is not obvious to me. Perhaps it's useful to mention this here to avoid confusion.
---
 docs/scripts.md | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/docs/scripts.md b/docs/scripts.md
index 0fbf18f1..b94690da 100644
--- a/docs/scripts.md
+++ b/docs/scripts.md
@@ -13,6 +13,8 @@ To train an SE-ResNet34 on ImageNet, locally distributed, 4 GPUs, one process pe
 `./distributed_train.sh 4 /data/imagenet --model seresnet34 --sched cosine --epochs 150 --warmup-epochs 5 --lr 0.4 --reprob 0.5 --remode pixel --batch-size 256 --amp -j 4`
 
 NOTE: It is recommended to use PyTorch 1.9+ w/ PyTorch native AMP and DDP instead of APEX AMP. `--amp` defaults to native AMP as of timm ver 0.4.3.  `--apex-amp` will force use of APEX components if they are installed.
+
+NOTE: For training and validation the same input size is used. Some models specify a different test input size which is only used in the validation script below.
  
 ## Validation / Inference Scripts
 
@@ -24,4 +26,4 @@ To validate with the model's pretrained weights (if they exist):
 
 To run inference from a checkpoint:
 
-`python inference.py /imagenet/validation/ --model mobilenetv3_large_100 --checkpoint ./output/train/model_best.pth.tar`
\ No newline at end of file
+`python inference.py /imagenet/validation/ --model mobilenetv3_large_100 --checkpoint ./output/train/model_best.pth.tar`