From ceae15042c110c5b0db2c16c4da053ee85b75226 Mon Sep 17 00:00:00 2001 From: Brian Vaughan Date: Wed, 14 Feb 2024 16:46:06 -0500 Subject: [PATCH] fix a type in code comment (#1063) was pointing to wrong class --- bitsandbytes/nn/modules.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/bitsandbytes/nn/modules.py b/bitsandbytes/nn/modules.py index 6eeecc273..2b7e1f067 100644 --- a/bitsandbytes/nn/modules.py +++ b/bitsandbytes/nn/modules.py @@ -275,7 +275,7 @@ class Linear4bit(nn.Linear): compute datatypes such as FP4 and NF4. In order to quantize a linear layer one should first load the original fp16 / bf16 weights into - the Linear8bitLt module, then call `quantized_module.to("cuda")` to quantize the fp16 / bf16 weights. + the Linear4bit module, then call `quantized_module.to("cuda")` to quantize the fp16 / bf16 weights. Example: