From 05e38f87dd5feb6b5e039744447b1cc9cfdb81e3 Mon Sep 17 00:00:00 2001 From: huangzhuang_3rd Date: Mon, 1 Jun 2026 18:31:43 +0800 Subject: [PATCH] =?UTF-8?q?v1.8.0:=20=E6=A8=A1=E5=9D=97=E5=8C=96=E9=87=8D?= =?UTF-8?q?=E6=9E=84=20+=20=E5=90=8E=E5=8F=B0=E7=99=BB=E5=BD=95=E8=AE=A4?= =?UTF-8?q?=E8=AF=81?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit --- templates/tools.html | 206 +++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 206 insertions(+) create mode 100644 templates/tools.html diff --git a/templates/tools.html b/templates/tools.html new file mode 100644 index 0000000..03bfa28 --- /dev/null +++ b/templates/tools.html @@ -0,0 +1,206 @@ + + + + + + ParamHub - 参数百科 + + + + + + + + +
+
+

+ + 实用工具 +

+

AI开发常用计算工具

+
+ + +
+

+ + 显存计算器 +

+

计算大模型所需的显存大小,并推荐合适的GPU

+ +
+
+ + +
+
+ + +
+
+ +
+
+ + + +
+ + +
+

+ + 精度与显存关系 +

+
+
+
FP32
+
4字节
+
全精度,最高精度
+
+
+
FP16
+
2字节
+
半精度,常用
+
+
+
INT8
+
1字节
+
8位量化
+
+
+
INT4
+
0.5字节
+
4位量化,最小
+
+
+
+ + +
+

+ + 常见模型显存需求参考 (FP16) +

+ + + + + + + + + + + + + + + + + +
模型参数量模型权重推荐显存推荐GPU
Mistral 7B7B14GB18GBRTX 3090/4090
Llama 2 13B13B26GB34GBRTX 4090 + 8GB
Llama 2 70B70B140GB182GBA100 80GB x2
Mixtral 8x7B47B94GB122GBA100 80GB x2
Qwen 72B72B144GB187GBH100 80GB x3
+
+
+ + + + \ No newline at end of file