Checkpoint Loader¶
Documentation¶
- Class name:
AV_CheckpointLoader
- Category:
Art Venture/Loaders
- Output node:
False
The AV_CheckpointLoader node is designed for loading model checkpoints with optional overrides for specific components such as the checkpoint itself, VAE, or LoRA models. It extends the functionality of a standard checkpoint loader by allowing users to specify alternative sources for model components, enhancing flexibility in model configuration and experimentation.
Input types¶
Required¶
ckpt_name
- Specifies the name of the checkpoint to load. This parameter can be overridden to load a different checkpoint if desired, providing flexibility in model experimentation.
- Comfy dtype:
COMBO[STRING]
- Python dtype:
str
vae_name
- Specifies the name of the VAE model to load. This parameter can be overridden to load a different VAE model if desired, allowing for experimentation with different VAE configurations.
- Comfy dtype:
COMBO[STRING]
- Python dtype:
str
clip_skip
- Indicates whether to skip loading the CLIP model. This parameter allows for selective loading of model components based on requirements.
- Comfy dtype:
INT
- Python dtype:
bool
lora_name
- Specifies the name of the LoRA model to load. This parameter can be overridden to load a different LoRA model if desired, enabling customization of the LoRA component.
- Comfy dtype:
COMBO[STRING]
- Python dtype:
str
lora_model_strength
- unknown
- Comfy dtype:
FLOAT
- Python dtype:
unknown
lora_clip_strength
- unknown
- Comfy dtype:
FLOAT
- Python dtype:
unknown
positive
- unknown
- Comfy dtype:
STRING
- Python dtype:
unknown
negative
- unknown
- Comfy dtype:
STRING
- Python dtype:
unknown
token_normalization
- unknown
- Comfy dtype:
COMBO[STRING]
- Python dtype:
unknown
weight_interpretation
- unknown
- Comfy dtype:
COMBO[STRING]
- Python dtype:
unknown
empty_latent_width
- unknown
- Comfy dtype:
INT
- Python dtype:
unknown
empty_latent_height
- unknown
- Comfy dtype:
INT
- Python dtype:
unknown
batch_size
- unknown
- Comfy dtype:
INT
- Python dtype:
unknown
Optional¶
lora_stack
- unknown
- Comfy dtype:
LORA_STACK
- Python dtype:
unknown
cnet_stack
- unknown
- Comfy dtype:
CONTROL_NET_STACK
- Python dtype:
unknown
ckpt_override
- unknown
- Comfy dtype:
STRING
- Python dtype:
unknown
vae_override
- unknown
- Comfy dtype:
STRING
- Python dtype:
unknown
lora_override
- unknown
- Comfy dtype:
STRING
- Python dtype:
unknown
Output types¶
MODEL
- Comfy dtype:
MODEL
- The loaded model instance.
- Python dtype:
torch.nn.Module
- Comfy dtype:
CONDITIONING+
- Comfy dtype:
CONDITIONING
- Positive conditioning components loaded or configured during the checkpoint loading process.
- Python dtype:
Dict[str, torch.Tensor]
- Comfy dtype:
CONDITIONING-
- Comfy dtype:
CONDITIONING
- Negative conditioning components loaded or configured during the checkpoint loading process.
- Python dtype:
Dict[str, torch.Tensor]
- Comfy dtype:
LATENT
- Comfy dtype:
LATENT
- Latent representations or configurations loaded or derived from the checkpoint.
- Python dtype:
torch.Tensor
- Comfy dtype:
VAE
- Comfy dtype:
VAE
- The VAE model loaded as part of the checkpoint, if applicable.
- Python dtype:
torch.nn.Module
- Comfy dtype:
CLIP
- Comfy dtype:
CLIP
- The CLIP model loaded as part of the checkpoint, if applicable.
- Python dtype:
torch.nn.Module
- Comfy dtype:
DEPENDENCIES
- Comfy dtype:
DEPENDENCIES
- Any additional dependencies or components loaded alongside the main model components.
- Python dtype:
List[torch.nn.Module]
- Comfy dtype:
Usage tips¶
- Infra type:
CPU
- Common nodes:
Source code¶
class AVCheckpointLoader(TSC_EfficientLoader):
@classmethod
def INPUT_TYPES(cls):
inputs = TSC_EfficientLoader.INPUT_TYPES()
inputs["optional"]["ckpt_override"] = ("STRING", {"default": "None"})
inputs["optional"]["vae_override"] = ("STRING", {"default": "None"})
inputs["optional"]["lora_override"] = ("STRING", {"default": "None"})
return inputs
CATEGORY = "Art Venture/Loaders"
def efficientloader(
self,
ckpt_name,
vae_name,
clip_skip,
lora_name,
*args,
ckpt_override="None",
vae_override="None",
lora_override="None",
**kwargs
):
if ckpt_override != "None":
ckpt_name = ckpt_override
if vae_override != "None":
vae_name = vae_override
if lora_override != "None":
lora_name = lora_override
return super().efficientloader(
ckpt_name, vae_name, clip_skip, lora_name, *args, **kwargs
)