Commit 15619e7

sayakpaulstevhliu

and

authored

Apply suggestions from code review

Co-authored-by: Steven Liu <59462357+stevhliu@users.noreply.github.com>

1 parent 6a8b3c9 commit 15619e7Copy full SHA for 15619e7

File tree

-1

lines changed

-1

lines changed

Lines changed: 2 additions & 1 deletion

Original file line number	Diff line number	Diff line change
@@ -36,6 +36,8 @@ Initialize [`~quantizers.PipelineQuantizationConfig`] with the following paramet
`36`	`36`
`37`	`37`	- `components_to_quantize` specifies which component(s) of the pipeline to quantize. Typically, you should quantize the most compute intensive components like the transformer. The text encoder is another component to consider quantizing if a pipeline has more than one such as [`FluxPipeline`]. The example below quantizes the T5 text encoder in [`FluxPipeline`] while keeping the CLIP model intact.
`38`	`38`
	`39`	+ `components_to_quantize` accepts either a list for multiple models or a string for a single model.
	`40`	`+`
`39`	`41`	The example below loads the bitsandbytes backend with the following arguments from [`~quantizers.quantization_config.BitsAndBytesConfig`], `load_in_4bit`, `bnb_4bit_quant_type`, and `bnb_4bit_compute_dtype`.
`40`	`42`
`41`	`43`	```py
`@@ -62,7 +64,6 @@ pipe = DiffusionPipeline.from_pretrained(`
`62`	`64`	`image = pipe("photo of a cute dog").images[0]`
`63`	`65`	```
`64`	`66`
`65`		-`components_to_quantize` doesn't have to be a list. You can also pass: `components_to_quantize="transformer"`.
`66`	`67`
`67`	`68`	`### Advanced quantization`
`68`	`69`

Comments

(0)