From ee5e4e9929086fba200b79f84ae7eff9c1bd5ab1 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Long=20M=2E=20L=C6=B0u?=
 <55311435+minhlong94@users.noreply.github.com>
Date: Sat, 21 Nov 2020 18:23:04 +0700
Subject: [PATCH] Update vit_pytorch.py

---
 vit_pytorch/vit_pytorch.py | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/vit_pytorch/vit_pytorch.py b/vit_pytorch/vit_pytorch.py
index 1520fe3..3bc2280 100644
--- a/vit_pytorch/vit_pytorch.py
+++ b/vit_pytorch/vit_pytorch.py
@@ -85,10 +85,10 @@ class Transformer(nn.Module):
 class ViT(nn.Module):
     def __init__(self, *, image_size, patch_size, num_classes, dim, depth, heads, mlp_dim, channels = 3, dropout = 0., emb_dropout = 0.):
         super().__init__()
-        assert image_size % patch_size == 0, 'image dimensions must be divisible by the patch size'
+        assert image_size % patch_size == 0, 'Image dimensions must be divisible by the patch size.'
         num_patches = (image_size // patch_size) ** 2
         patch_dim = channels * patch_size ** 2
-        assert num_patches > MIN_NUM_PATCHES, f'your number of patches ({num_patches}) is way too small for attention to be effective. try decreasing your patch size'
+        assert num_patches > MIN_NUM_PATCHES, f'your number of patches ({num_patches}) is way too small for attention to be effective (at least 16). Try decreasing your patch size'
 
         self.patch_size = patch_size