src/pixparse/task/task_cruller_eval_docvqa.py [132:155]:
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
            self.model.text_decoder.trunk.resize_token_embeddings(
                len(self.tokenizer.trunk)
            )

# ------ 
        self.loss = nn.CrossEntropyLoss(ignore_index=-100)
        self.has_no_sync = False
        self.num_image_chs = 1 if cfg.model.image_encoder.image_fmt == "L" else 3

        # TODO refactor, used in many tasks

        img_mean = self.model.image_encoder.trunk.pretrained_cfg["mean"]
        img_std = self.model.image_encoder.trunk.pretrained_cfg["std"]

        self.img_mean = (
            sum(img_mean) / len(img_mean)
            if cfg.model.image_encoder.image_fmt == "L"
            else img_mean
        )
        self.img_std = (
            sum(img_std) / len(img_std)
            if cfg.model.image_encoder.image_fmt == "L"
            else img_std
        )
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -



src/pixparse/task/task_cruller_finetune_docvqa.py [132:153]:
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
            self.model.text_decoder.trunk.resize_token_embeddings(
                len(self.tokenizer.trunk)
            )

        self.loss = nn.CrossEntropyLoss(ignore_index=-100)
        self.has_no_sync = False
        self.num_image_chs = 1 if cfg.model.image_encoder.image_fmt == "L" else 3

        img_mean = self.model.image_encoder.trunk.pretrained_cfg["mean"]
        img_std = self.model.image_encoder.trunk.pretrained_cfg["std"]

        self.img_mean = (
            sum(img_mean) / len(img_mean)
            if cfg.model.image_encoder.image_fmt == "L"
            else img_mean
        )

        self.img_std = (
            sum(img_std) / len(img_std)
            if cfg.model.image_encoder.image_fmt == "L"
            else img_std
        )
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -