Fix documentation for FluxPipeline (#10563)

Fix argument name in 8bit quantized example

Found a tiny mistake in the documentation where the text encoder model was passed to the wrong argument in the FluxPipeline.from_pretrained function.

Author: ohm314

docs/source/en/api/pipelines/flux.md

@@ -367,7 +367,7 @@ transformer_8bit = FluxTransformer2DModel.from_pretrained(
 
 pipeline = FluxPipeline.from_pretrained(
     "black-forest-labs/FLUX.1-dev",
-    text_encoder=text_encoder_8bit,
+    text_encoder_2=text_encoder_8bit,
     transformer=transformer_8bit,
     torch_dtype=torch.float16,
     device_map="balanced",