comfyui-bert/__init__.py

from .cosmos_modernbert import CosmosBert
from comfy.comfy_types import IO
import torch
from transformers import AutoTokenizer
from transformers.models.modernbert.configuration_modernbert import ModernBertConfig

class CosmosBertTextEncode:
    @classmethod
    def INPUT_TYPES(s):
        return {
            "required": {
                "text": (IO.STRING, {"multiline": True, "dynamicPrompts": True, "tooltip": "The text to be encoded."}),
                "clip": (IO.CLIP, {"tooltip": "The CLIP model used for encoding the text."})
            }
        }
    RETURN_TYPES = (IO.CONDITIONING,)
    OUTPUT_TOOLTIPS = ("A conditioning containing the embedded text used to guide the diffusion model.",)
    FUNCTION = "encode"

    CATEGORY = "conditioning"
    DESCRIPTION = "Encodes a text prompt using a CLIP model into an embedding that can be used to guide the diffusion model towards generating specific images."
    SEARCH_ALIASES = ["text", "prompt", "text prompt", "positive prompt", "negative prompt", "encode text", "text encoder", "encode prompt"]

    def encode(self, clip, text):
        bert = CosmosBert(ModernBertConfig())
        bert.load_safetensors('model.safetensors')
        if clip is None:
            raise RuntimeError("ERROR: clip input is invalid: None\n\nIf the clip is from a checkpoint loader node your checkpoint does not contain a valid clip or text encoder model.")
        tokenizer = AutoTokenizer.from_pretrained('nightknocker/cosmos-bert')
        inputs = tokenizer(text, return_tensors='pt')
        crossattn_emb, keys = bert.encode(**inputs)

        return [(crossattn_emb, keys)],

NODE_CLASS_MAPPINGS = {
    "CosmosBertTextEncode": CosmosBertTextEncode,
}