stable-diffusion-webui/modules/sd_hijack_clip.py

import math
from collections import namedtuple

import torch

from modules import prompt_parser, devices, sd_hijack, sd_emphasis
from modules.shared import opts


class PromptChunk:
    """
    This object contains token ids, weight (multipliers:1.4) and textual inversion embedding info for a chunk of prompt.
    If a prompt is short, it is represented by one PromptChunk, otherwise, multiple are necessary.
    Each PromptChunk contains an exact amount of tokens - 77, which includes one for start and end token,
    so just 75 tokens from prompt.
    """

    def __init__(self):
        self.tokens = []
        self.multipliers = []
        self.fixes = []


PromptChunkFix = namedtuple('PromptChunkFix', ['offset', 'embedding'])
"""An object of this type is a marker showing that textual inversion embedding's vectors have to placed at offset in the prompt
chunk. Those objects are found in PromptChunk.fixes and, are placed into FrozenCLIPEmbedderWithCustomWordsBase.hijack.fixes, and finally
are applied by sd_hijack.EmbeddingsWithFixes's forward function."""


class TextConditionalModel(torch.nn.Module):
    def __init__(self):
        super().__init__()

        self.hijack = sd_hijack.model_hijack
        self.chunk_length = 75

        self.is_trainable = False
        self.input_key = 'txt'
        self.return_pooled = False

        self.comma_token = None
        self.id_start = None
        self.id_end = None
        self.id_pad = None

    def empty_chunk(self):
        """creates an empty PromptChunk and returns it"""

        chunk = PromptChunk()
        chunk.tokens = [self.id_start] + [self.id_end] * (self.chunk_length + 1)
        chunk.multipliers = [1.0] * (self.chunk_length + 2)
        return chunk

    def get_target_prompt_token_count(self, token_count):
        """returns the maximum number of tokens a prompt of a known length can have before it requires one more PromptChunk to be represented"""

        return math.ceil(max(token_count, 1) / self.chunk_length) * self.chunk_length

    def tokenize(self, texts):
        """Converts a batch of texts into a batch of token ids"""

        raise NotImplementedError

    def encode_with_transformers(self, tokens):
        """
        converts a batch of token ids (in python lists) into a single tensor with numeric representation of those tokens;
        All python lists with tokens are assumed to have same length, usually 77.
        if input is a list with B elements and each element has T tokens, expected output shape is (B, T, C), where C depends on
        model - can be 768 and 1024.
        Among other things, this call will read self.hijack.fixes, apply it to its inputs, and clear it (setting it to None).
        """

        raise NotImplementedError

    def encode_embedding_init_text(self, init_text, nvpt):
        """Converts text into a tensor with this text's tokens' embeddings. Note that those are embeddings before they are passed through
        transformers. nvpt is used as a maximum length in tokens. If text produces less teokens than nvpt, only this many is returned."""

        raise NotImplementedError

    def tokenize_line(self, line):
        """
        this transforms a single prompt into a list of PromptChunk objects - as many as needed to
        represent the prompt.
        Returns the list and the total number of tokens in the prompt.
        """

        if opts.emphasis != "None":
            parsed = prompt_parser.parse_prompt_attention(line)
        else:
            parsed = [[line, 1.0]]

        tokenized = self.tokenize([text for text, _ in parsed])

        chunks = []
        chunk = PromptChunk()
        token_count = 0
        last_comma = -1

        def next_chunk(is_last=False):
            """puts current chunk into the list of results and produces the next one - empty;
            if is_last is true, tokens <end-of-text> tokens at the end won't add to token_count"""
            nonlocal token_count
            nonlocal last_comma
            nonlocal chunk

            if is_last:
                token_count += len(chunk.tokens)
            else:
                token_count += self.chunk_length

            to_add = self.chunk_length - len(chunk.tokens)
            if to_add > 0:
                chunk.tokens += [self.id_end] * to_add
                chunk.multipliers += [1.0] * to_add

            chunk.tokens = [self.id_start] + chunk.tokens + [self.id_end]
            chunk.multipliers = [1.0] + chunk.multipliers + [1.0]

            last_comma = -1
            chunks.append(chunk)
            chunk = PromptChunk()

        for tokens, (text, weight) in zip(tokenized, parsed):
            if text == 'BREAK' and weight == -1:
                next_chunk()
                continue

            position = 0
            while position < len(tokens):
                token = tokens[position]

                if token == self.comma_token:
                    last_comma = len(chunk.tokens)

                # this is when we are at the end of allotted 75 tokens for the current chunk, and the current token is not a comma. opts.comma_padding_backtrack
                # is a setting that specifies that if there is a comma nearby, the text after the comma should be moved out of this chunk and into the next.
                elif opts.comma_padding_backtrack != 0 and len(chunk.tokens) == self.chunk_length and last_comma != -1 and len(chunk.tokens) - last_comma <= opts.comma_padding_backtrack:
                    break_location = last_comma + 1

                    reloc_tokens = chunk.tokens[break_location:]
                    reloc_mults = chunk.multipliers[break_location:]

                    chunk.tokens = chunk.tokens[:break_location]
                    chunk.multipliers = chunk.multipliers[:break_location]

                    next_chunk()
                    chunk.tokens = reloc_tokens
                    chunk.multipliers = reloc_mults

                if len(chunk.tokens) == self.chunk_length:
                    next_chunk()

                embedding, embedding_length_in_tokens = self.hijack.embedding_db.find_embedding_at_position(tokens, position)
                if embedding is None:
                    chunk.tokens.append(token)
                    chunk.multipliers.append(weight)
                    position += 1
                    continue

                emb_len = int(embedding.vectors)
                if len(chunk.tokens) + emb_len > self.chunk_length:
                    next_chunk()

                chunk.fixes.append(PromptChunkFix(len(chunk.tokens), embedding))

                chunk.tokens += [0] * emb_len
                chunk.multipliers += [weight] * emb_len
                position += embedding_length_in_tokens

        if chunk.tokens or not chunks:
            next_chunk(is_last=True)

        return chunks, token_count

    def process_texts(self, texts):
        """
        Accepts a list of texts and calls tokenize_line() on each, with cache. Returns the list of results and maximum
        length, in tokens, of all texts.
        """

        token_count = 0

        cache = {}
        batch_chunks = []
        for line in texts:
            if line in cache:
                chunks = cache[line]
            else:
                chunks, current_token_count = self.tokenize_line(line)
                token_count = max(current_token_count, token_count)

                cache[line] = chunks

            batch_chunks.append(chunks)

        return batch_chunks, token_count

    def forward(self, texts):
        """
        Accepts an array of texts; Passes texts through transformers network to create a tensor with numerical representation of those texts.
        Returns a tensor with shape of (B, T, C), where B is length of the array; T is length, in tokens, of texts (including padding) - T will
        be a multiple of 77; and C is dimensionality of each token - for SD1 it's 768, for SD2 it's 1024, and for SDXL it's 1280.
        An example shape returned by this function can be: (2, 77, 768).
        For SDXL, instead of returning one tensor avobe, it returns a tuple with two: the other one with shape (B, 1280) with pooled values.
        Webui usually sends just one text at a time through this function - the only time when texts is an array with more than one element
        is when you do prompt editing: "a picture of a [cat:dog:0.4] eating ice cream"
        """

        batch_chunks, token_count = self.process_texts(texts)

        used_embeddings = {}
        chunk_count = max([len(x) for x in batch_chunks])

        zs = []
        for i in range(chunk_count):
            batch_chunk = [chunks[i] if i < len(chunks) else self.empty_chunk() for chunks in batch_chunks]

            tokens = [x.tokens for x in batch_chunk]
            multipliers = [x.multipliers for x in batch_chunk]
            self.hijack.fixes = [x.fixes for x in batch_chunk]

            for fixes in self.hijack.fixes:
                for _position, embedding in fixes:
                    used_embeddings[embedding.name] = embedding
            devices.torch_npu_set_device()
            z = self.process_tokens(tokens, multipliers)
            zs.append(z)

        if opts.textual_inversion_add_hashes_to_infotext and used_embeddings:
            hashes = []
            for name, embedding in used_embeddings.items():
                shorthash = embedding.shorthash
                if not shorthash:
                    continue

                name = name.replace(":", "").replace(",", "")
                hashes.append(f"{name}: {shorthash}")

            if hashes:
                if self.hijack.extra_generation_params.get("TI hashes"):
                    hashes.append(self.hijack.extra_generation_params.get("TI hashes"))
                self.hijack.extra_generation_params["TI hashes"] = ", ".join(hashes)

        if any(x for x in texts if "(" in x or "[" in x) and opts.emphasis != "Original":
            self.hijack.extra_generation_params["Emphasis"] = opts.emphasis

        if self.return_pooled:
            return torch.hstack(zs), zs[0].pooled
        else:
            return torch.hstack(zs)

    def process_tokens(self, remade_batch_tokens, batch_multipliers):
        """
        sends one single prompt chunk to be encoded by transformers neural network.
        remade_batch_tokens is a batch of tokens - a list, where every element is a list of tokens; usually
        there are exactly 77 tokens in the list. batch_multipliers is the same but for multipliers instead of tokens.
        Multipliers are used to give more or less weight to the outputs of transformers network. Each multiplier
        corresponds to one token.
        """
        tokens = torch.asarray(remade_batch_tokens).to(devices.device)

        # this is for SD2: SD1 uses the same token for padding and end of text, while SD2 uses different ones.
        if self.id_end != self.id_pad:
            for batch_pos in range(len(remade_batch_tokens)):
                index = remade_batch_tokens[batch_pos].index(self.id_end)
                tokens[batch_pos, index+1:tokens.shape[1]] = self.id_pad

        z = self.encode_with_transformers(tokens)

        pooled = getattr(z, 'pooled', None)

        emphasis = sd_emphasis.get_current_option(opts.emphasis)()
        emphasis.tokens = remade_batch_tokens
        emphasis.multipliers = torch.asarray(batch_multipliers).to(devices.device)
        emphasis.z = z

        emphasis.after_transformers()

        z = emphasis.z

        if pooled is not None:
            z.pooled = pooled

        return z


class FrozenCLIPEmbedderWithCustomWordsBase(TextConditionalModel):
    """A pytorch module that is a wrapper for FrozenCLIPEmbedder module. it enhances FrozenCLIPEmbedder, making it possible to
    have unlimited prompt length and assign weights to tokens in prompt.
    """

    def __init__(self, wrapped, hijack):
        super().__init__()

        self.hijack = hijack

        self.wrapped = wrapped
        """Original FrozenCLIPEmbedder module; can also be FrozenOpenCLIPEmbedder or xlmr.BertSeriesModelWithTransformation,
        depending on model."""

        self.is_trainable = getattr(wrapped, 'is_trainable', False)
        self.input_key = getattr(wrapped, 'input_key', 'txt')
        self.return_pooled = getattr(self.wrapped, 'return_pooled', False)

        self.legacy_ucg_val = None  # for sgm codebase

    def forward(self, texts):
        if opts.use_old_emphasis_implementation:
            import modules.sd_hijack_clip_old
            return modules.sd_hijack_clip_old.forward_old(self, texts)

        return super().forward(texts)


class FrozenCLIPEmbedderWithCustomWords(FrozenCLIPEmbedderWithCustomWordsBase):
    def __init__(self, wrapped, hijack):
        super().__init__(wrapped, hijack)
        self.tokenizer = wrapped.tokenizer

        vocab = self.tokenizer.get_vocab()

        self.comma_token = vocab.get(',</w>', None)

        self.token_mults = {}
        tokens_with_parens = [(k, v) for k, v in vocab.items() if '(' in k or ')' in k or '[' in k or ']' in k]
        for text, ident in tokens_with_parens:
            mult = 1.0
            for c in text:
                if c == '[':
                    mult /= 1.1
                if c == ']':
                    mult *= 1.1
                if c == '(':
                    mult *= 1.1
                if c == ')':
                    mult /= 1.1

            if mult != 1.0:
                self.token_mults[ident] = mult

        self.id_start = self.wrapped.tokenizer.bos_token_id
        self.id_end = self.wrapped.tokenizer.eos_token_id
        self.id_pad = self.id_end

    def tokenize(self, texts):
        tokenized = self.wrapped.tokenizer(texts, truncation=False, add_special_tokens=False)["input_ids"]

        return tokenized

    def encode_with_transformers(self, tokens):
        outputs = self.wrapped.transformer(input_ids=tokens, output_hidden_states=-opts.CLIP_stop_at_last_layers)

        if opts.CLIP_stop_at_last_layers > 1:
            z = outputs.hidden_states[-opts.CLIP_stop_at_last_layers]
            z = self.wrapped.transformer.text_model.final_layer_norm(z)
        else:
            z = outputs.last_hidden_state

        return z

    def encode_embedding_init_text(self, init_text, nvpt):
        embedding_layer = self.wrapped.transformer.text_model.embeddings
        ids = self.wrapped.tokenizer(init_text, max_length=nvpt, return_tensors="pt", add_special_tokens=False)["input_ids"]
        embedded = embedding_layer.token_embedding.wrapped(ids.to(embedding_layer.token_embedding.wrapped.weight.device)).squeeze(0)

        return embedded


class FrozenCLIPEmbedderForSDXLWithCustomWords(FrozenCLIPEmbedderWithCustomWords):
    def __init__(self, wrapped, hijack):
        super().__init__(wrapped, hijack)

    def encode_with_transformers(self, tokens):
        outputs = self.wrapped.transformer(input_ids=tokens, output_hidden_states=self.wrapped.layer == "hidden")

        if opts.sdxl_clip_l_skip is True:
            z = outputs.hidden_states[-opts.CLIP_stop_at_last_layers]
        elif self.wrapped.layer == "last":
            z = outputs.last_hidden_state
        else:
            z = outputs.hidden_states[self.wrapped.layer_idx]

        return z