remove unused files

dialoggen/dialoggen_demo.py

@@ -1,189 +0,0 @@
-import argparse
-import torch
-import sys
-import os
-# 添加当前命令行运行的目录到 sys.path
-sys.path.append(os.getcwd()+"/dialoggen")
-
-
-from llava.constants import (
-    IMAGE_TOKEN_INDEX,
-    DEFAULT_IMAGE_TOKEN,
-    DEFAULT_IM_START_TOKEN,
-    DEFAULT_IM_END_TOKEN,
-    IMAGE_PLACEHOLDER,
-)
-from llava.conversation import conv_templates, SeparatorStyle
-from llava.model.builder import load_pretrained_model
-from llava.utils import disable_torch_init
-from llava.mm_utils import (
-    process_images,
-    tokenizer_image_token,
-    get_model_name_from_path,
-)
-
-import requests
-from PIL import Image
-from io import BytesIO
-import re
-
-
-def image_parser(image_file, sep=','):
-    out = image_file.split(sep)
-    return out
-
-
-def load_image(image_file):
-    if image_file.startswith("http") or image_file.startswith("https"):
-        response = requests.get(image_file)
-        image = Image.open(BytesIO(response.content)).convert("RGB")
-    else:
-        image = Image.open(image_file).convert("RGB")
-    return image
-
-
-def load_images(image_files):
-    out = []
-    for image_file in image_files:
-        image = load_image(image_file)
-        out.append(image)
-    return out
-
-
-def init_dialoggen_model(model_path, model_base=None, load_4bit=False):
-    model_name = get_model_name_from_path(model_path)
-    tokenizer, model, image_processor, context_len = load_pretrained_model(
-        model_path, model_base, model_name, llava_type_model=True, load_4bit=load_4bit)
-    return {"tokenizer": tokenizer,
-            "model": model,
-            "image_processor": image_processor}
-
-
-def eval_model(models,
-               query='详细描述一下这张图片',
-               image_file=None,
-               sep=',',
-               temperature=0.2,
-               top_p=None,
-               num_beams=1,
-               max_new_tokens=512,
-               return_history=False,
-               history=None,
-               skip_special=False
-               ):
-    # Model
-    disable_torch_init()
-
-    qs = query
-    image_token_se = DEFAULT_IM_START_TOKEN + DEFAULT_IMAGE_TOKEN + DEFAULT_IM_END_TOKEN
-    if IMAGE_PLACEHOLDER in qs:
-        if models["model"].config.mm_use_im_start_end:
-            qs = re.sub(IMAGE_PLACEHOLDER, image_token_se, qs)
-        else:
-            qs = re.sub(IMAGE_PLACEHOLDER, DEFAULT_IMAGE_TOKEN, qs)
-    else:
-        if models["model"].config.mm_use_im_start_end:
-            qs = image_token_se + "\n" + qs
-        else:
-            qs = DEFAULT_IMAGE_TOKEN + "\n" + qs
-            
-    if not history:
-        conv = conv_templates['llava_v1'].copy()
-    else:
-        conv = history
-
-    if skip_special:
-        conv.append_message(conv.roles[0], query)
-    else:
-        conv.append_message(conv.roles[0], qs)
-    conv.append_message(conv.roles[1], None)
-    prompt = conv.get_prompt()
-
-    if image_file is not None:
-        image_files = image_parser(image_file, sep=sep)
-        images = load_images(image_files)
-        image_sizes = [x.size for x in images]
-        images_tensor = process_images(
-            images,
-            models["image_processor"],
-            models["model"].config
-        ).to(models["model"].device, dtype=torch.float16)
-    else:
-        # fomatted input as training data
-        image_sizes = [(1024, 1024)]
-        images_tensor = torch.zeros(1, 5, 3, models["image_processor"].crop_size["height"], models["image_processor"].crop_size["width"])
-        images_tensor = images_tensor.to(models["model"].device, dtype=torch.float16)
-
-    input_ids = (
-        tokenizer_image_token(prompt, models["tokenizer"], IMAGE_TOKEN_INDEX, return_tensors="pt")
-        .unsqueeze(0)
-        .cuda()
-    )
-    with torch.inference_mode():
-        output_ids = models["model"].generate(
-            input_ids,
-            images=images_tensor,
-            image_sizes=image_sizes,
-            do_sample=True if temperature > 0 else False,
-            temperature=temperature,
-            top_p=top_p,
-            num_beams=num_beams,
-            max_new_tokens=max_new_tokens,
-            use_cache=True,
-        )
-
-    outputs = models["tokenizer"].batch_decode(output_ids, skip_special_tokens=True)[0].strip()
-    if return_history:
-        return outputs, conv
-    return outputs
-
-
-def remove_prefix(text):
-    if text.startswith("<画图>"):
-        return text[len("<画图>"):], True
-    elif text.startswith("对不起"):
-        # 拒绝画图
-        return "", False
-    else:
-        return text, True
-
-
-class DialogGen(object):
-    def __init__(self, model_path, load_4bit=False):
-        self.models = init_dialoggen_model(model_path, load_4bit=load_4bit)
-        self.query_template = "请先判断用户的意图，若为画图则在输出前加入<画图>:{}"
-
-    def __call__(self, prompt, return_history=False, history=None, skip_special=False):
-        enhanced_prompt = eval_model(
-            models=self.models,
-            query=self.query_template.format(prompt),
-            image_file=None,
-            return_history=return_history,
-            history=history,
-            skip_special=skip_special
-        )
-        if return_history:
-            return enhanced_prompt
-
-        enhanced_prompt, compliance = remove_prefix(enhanced_prompt)
-        if not compliance:
-            return False, ""
-        return True, enhanced_prompt
-
-
-if __name__ == "__main__":
-    parser = argparse.ArgumentParser()
-    parser.add_argument('--model_path', type=str, default='./ckpts/dialoggen')
-    parser.add_argument('--prompt', type=str, default='画一只小猫')
-    parser.add_argument('--image_file', type=str, default=None) # 'images/demo1.jpeg'
-    args = parser.parse_args()
-
-    query = f"请先判断用户的意图，若为画图则在输出前加入<画图>:{args.prompt}"
-
-    models = init_dialoggen_model(args.model_path)
-
-    res = eval_model(models,
-        query=query,
-        image_file=args.image_file,
-    )
-    print(res)

dialoggen/llava/__init__.py

@@ -1 +0,0 @@
-from .model import LlavaLlamaForCausalLM

dialoggen/llava/constants.py

@@ -1,13 +0,0 @@
-CONTROLLER_HEART_BEAT_EXPIRATION = 30
-WORKER_HEART_BEAT_INTERVAL = 15
-
-LOGDIR = "."
-
-# Model Constants
-IGNORE_INDEX = -100
-IMAGE_TOKEN_INDEX = -200
-DEFAULT_IMAGE_TOKEN = "<image>"
-DEFAULT_IMAGE_PATCH_TOKEN = "<im_patch>"
-DEFAULT_IM_START_TOKEN = "<im_start>"
-DEFAULT_IM_END_TOKEN = "<im_end>"
-IMAGE_PLACEHOLDER = "<image-placeholder>"

dialoggen/llava/conversation.py

@@ -1,396 +0,0 @@
-import dataclasses
-from enum import auto, Enum
-from typing import List, Tuple
-import base64
-from io import BytesIO
-from PIL import Image
-
-
-class SeparatorStyle(Enum):
-    """Different separator style."""
-    SINGLE = auto()
-    TWO = auto()
-    MPT = auto()
-    PLAIN = auto()
-    LLAMA_2 = auto()
-
-
-@dataclasses.dataclass
-class Conversation:
-    """A class that keeps all conversation history."""
-    system: str
-    roles: List[str]
-    messages: List[List[str]]
-    offset: int
-    sep_style: SeparatorStyle = SeparatorStyle.SINGLE
-    sep: str = "###"
-    sep2: str = None
-    version: str = "Unknown"
-
-    skip_next: bool = False
-
-    def get_prompt(self):
-        messages = self.messages
-        if len(messages) > 0 and type(messages[0][1]) is tuple:
-            messages = self.messages.copy()
-            init_role, init_msg = messages[0].copy()
-            init_msg = init_msg[0].replace("<image>", "").strip()
-            if 'mmtag' in self.version:
-                messages[0] = (init_role, init_msg)
-                messages.insert(0, (self.roles[0], "<Image><image></Image>"))
-                messages.insert(1, (self.roles[1], "Received."))
-            else:
-                messages[0] = (init_role, "<image>\n" + init_msg)
-
-        if self.sep_style == SeparatorStyle.SINGLE:
-            ret = self.system + self.sep
-            for role, message in messages:
-                if message:
-                    if type(message) is tuple:
-                        message, _, _ = message
-                    ret += role + ": " + message + self.sep
-                else:
-                    ret += role + ":"
-        elif self.sep_style == SeparatorStyle.TWO:
-            seps = [self.sep, self.sep2]
-            ret = self.system + seps[0]
-            for i, (role, message) in enumerate(messages):
-                if message:
-                    if type(message) is tuple:
-                        message, _, _ = message
-                    ret += role + ": " + message + seps[i % 2]
-                else:
-                    ret += role + ":"
-        elif self.sep_style == SeparatorStyle.MPT:
-            ret = self.system + self.sep
-            for role, message in messages:
-                if message:
-                    if type(message) is tuple:
-                        message, _, _ = message
-                    ret += role + message + self.sep
-                else:
-                    ret += role
-        elif self.sep_style == SeparatorStyle.LLAMA_2:
-            wrap_sys = lambda msg: f"<<SYS>>\n{msg}\n<</SYS>>\n\n" if len(msg) > 0 else msg
-            wrap_inst = lambda msg: f"[INST] {msg} [/INST]"
-            ret = ""
-
-            for i, (role, message) in enumerate(messages):
-                if i == 0:
-                    assert message, "first message should not be none"
-                    assert role == self.roles[0], "first message should come from user"
-                if message:
-                    if type(message) is tuple:
-                        message, _, _ = message
-                    if i == 0: message = wrap_sys(self.system) + message
-                    if i % 2 == 0:
-                        message = wrap_inst(message)
-                        ret += self.sep + message
-                    else:
-                        ret += " " + message + " " + self.sep2
-                else:
-                    ret += ""
-            ret = ret.lstrip(self.sep)
-        elif self.sep_style == SeparatorStyle.PLAIN:
-            seps = [self.sep, self.sep2]
-            ret = self.system
-            for i, (role, message) in enumerate(messages):
-                if message:
-                    if type(message) is tuple:
-                        message, _, _ = message
-                    ret += message + seps[i % 2]
-                else:
-                    ret += ""
-        else:
-            raise ValueError(f"Invalid style: {self.sep_style}")
-
-        return ret
-
-    def append_message(self, role, message):
-        self.messages.append([role, message])
-
-    def process_image(self, image, image_process_mode, return_pil=False, image_format='PNG', max_len=1344, min_len=672):
-        if image_process_mode == "Pad":
-            def expand2square(pil_img, background_color=(122, 116, 104)):
-                width, height = pil_img.size
-                if width == height:
-                    return pil_img
-                elif width > height:
-                    result = Image.new(pil_img.mode, (width, width), background_color)
-                    result.paste(pil_img, (0, (width - height) // 2))
-                    return result
-                else:
-                    result = Image.new(pil_img.mode, (height, height), background_color)
-                    result.paste(pil_img, ((height - width) // 2, 0))
-                    return result
-            image = expand2square(image)
-        elif image_process_mode in ["Default", "Crop"]:
-            pass
-        elif image_process_mode == "Resize":
-            image = image.resize((336, 336))
-        else:
-            raise ValueError(f"Invalid image_process_mode: {image_process_mode}")
-        if max(image.size) > max_len:
-            max_hw, min_hw = max(image.size), min(image.size)
-            aspect_ratio = max_hw / min_hw
-            shortest_edge = int(min(max_len / aspect_ratio, min_len, min_hw))
-            longest_edge = int(shortest_edge * aspect_ratio)
-            W, H = image.size
-            if H > W:
-                H, W = longest_edge, shortest_edge
-            else:
-                H, W = shortest_edge, longest_edge
-            image = image.resize((W, H))
-        if return_pil:
-            return image
-        else:
-            buffered = BytesIO()
-            image.save(buffered, format=image_format)
-            img_b64_str = base64.b64encode(buffered.getvalue()).decode()
-            return img_b64_str
-
-    def get_images(self, return_pil=False):
-        images = []
-        for i, (role, msg) in enumerate(self.messages[self.offset:]):
-            if i % 2 == 0:
-                if type(msg) is tuple:
-                    msg, image, image_process_mode = msg
-                    image = self.process_image(image, image_process_mode, return_pil=return_pil)
-                    images.append(image)
-        return images
-
-    def to_gradio_chatbot(self):
-        ret = []
-        for i, (role, msg) in enumerate(self.messages[self.offset:]):
-            if i % 2 == 0:
-                if type(msg) is tuple:
-                    msg, image, image_process_mode = msg
-                    img_b64_str = self.process_image(
-                        image, "Default", return_pil=False,
-                        image_format='JPEG')
-                    img_str = f'<img src="data:image/jpeg;base64,{img_b64_str}" alt="user upload image" />'
-                    msg = img_str + msg.replace('<image>', '').strip()
-                    ret.append([msg, None])
-                else:
-                    ret.append([msg, None])
-            else:
-                ret[-1][-1] = msg
-        return ret
-
-    def copy(self):
-        return Conversation(
-            system=self.system,
-            roles=self.roles,
-            messages=[[x, y] for x, y in self.messages],
-            offset=self.offset,
-            sep_style=self.sep_style,
-            sep=self.sep,
-            sep2=self.sep2,
-            version=self.version)
-
-    def dict(self):
-        if len(self.get_images()) > 0:
-            return {
-                "system": self.system,
-                "roles": self.roles,
-                "messages": [[x, y[0] if type(y) is tuple else y] for x, y in self.messages],
-                "offset": self.offset,
-                "sep": self.sep,
-                "sep2": self.sep2,
-            }
-        return {
-            "system": self.system,
-            "roles": self.roles,
-            "messages": self.messages,
-            "offset": self.offset,
-            "sep": self.sep,
-            "sep2": self.sep2,
-        }
-
-
-conv_vicuna_v0 = Conversation(
-    system="A chat between a curious human and an artificial intelligence assistant. "
-           "The assistant gives helpful, detailed, and polite answers to the human's questions.",
-    roles=("Human", "Assistant"),
-    messages=(
-        ("Human", "What are the key differences between renewable and non-renewable energy sources?"),
-        ("Assistant",
-            "Renewable energy sources are those that can be replenished naturally in a relatively "
-            "short amount of time, such as solar, wind, hydro, geothermal, and biomass. "
-            "Non-renewable energy sources, on the other hand, are finite and will eventually be "
-            "depleted, such as coal, oil, and natural gas. Here are some key differences between "
-            "renewable and non-renewable energy sources:\n"
-            "1. Availability: Renewable energy sources are virtually inexhaustible, while non-renewable "
-            "energy sources are finite and will eventually run out.\n"
-            "2. Environmental impact: Renewable energy sources have a much lower environmental impact "
-            "than non-renewable sources, which can lead to air and water pollution, greenhouse gas emissions, "
-            "and other negative effects.\n"
-            "3. Cost: Renewable energy sources can be more expensive to initially set up, but they typically "
-            "have lower operational costs than non-renewable sources.\n"
-            "4. Reliability: Renewable energy sources are often more reliable and can be used in more remote "
-            "locations than non-renewable sources.\n"
-            "5. Flexibility: Renewable energy sources are often more flexible and can be adapted to different "
-            "situations and needs, while non-renewable sources are more rigid and inflexible.\n"
-            "6. Sustainability: Renewable energy sources are more sustainable over the long term, while "
-            "non-renewable sources are not, and their depletion can lead to economic and social instability.\n")
-    ),
-    offset=2,
-    sep_style=SeparatorStyle.SINGLE,
-    sep="###",
-)
-
-conv_vicuna_v1 = Conversation(
-    system="A chat between a curious user and an artificial intelligence assistant. "
-    "The assistant gives helpful, detailed, and polite answers to the user's questions.",
-    roles=("USER", "ASSISTANT"),
-    version="v1",
-    messages=(),
-    offset=0,
-    sep_style=SeparatorStyle.TWO,
-    sep=" ",
-    sep2="</s>",
-)
-
-conv_llama_2 = Conversation(
-    system="""You are a helpful, respectful and honest assistant. Always answer as helpfully as possible, while being safe.  Your answers should not include any harmful, unethical, racist, sexist, toxic, dangerous, or illegal content. Please ensure that your responses are socially unbiased and positive in nature.
-
-If a question does not make any sense, or is not factually coherent, explain why instead of answering something not correct. If you don't know the answer to a question, please don't share false information.""",
-    roles=("USER", "ASSISTANT"),
-    version="llama_v2",
-    messages=(),
-    offset=0,
-    sep_style=SeparatorStyle.LLAMA_2,
-    sep="<s>",
-    sep2="</s>",
-)
-
-conv_llava_llama_2 = Conversation(
-    system="You are a helpful language and vision assistant. "
-           "You are able to understand the visual content that the user provides, "
-           "and assist the user with a variety of tasks using natural language.",
-    roles=("USER", "ASSISTANT"),
-    version="llama_v2",
-    messages=(),
-    offset=0,
-    sep_style=SeparatorStyle.LLAMA_2,
-    sep="<s>",
-    sep2="</s>",
-)
-
-conv_mpt = Conversation(
-    system="""<|im_start|>system
-A conversation between a user and an LLM-based AI assistant. The assistant gives helpful and honest answers.""",
-    roles=("<|im_start|>user\n", "<|im_start|>assistant\n"),
-    version="mpt",
-    messages=(),
-    offset=0,
-    sep_style=SeparatorStyle.MPT,
-    sep="<|im_end|>",
-)
-
-conv_llava_plain = Conversation(
-    system="",
-    roles=("", ""),
-    messages=(
-    ),
-    offset=0,
-    sep_style=SeparatorStyle.PLAIN,
-    sep="\n",
-)
-
-conv_llava_v0 = Conversation(
-    system="A chat between a curious human and an artificial intelligence assistant. "
-           "The assistant gives helpful, detailed, and polite answers to the human's questions.",
-    roles=("Human", "Assistant"),
-    messages=(
-    ),
-    offset=0,
-    sep_style=SeparatorStyle.SINGLE,
-    sep="###",
-)
-
-conv_llava_v0_mmtag = Conversation(
-    system="A chat between a curious user and an artificial intelligence assistant. "
-           "The assistant is able to understand the visual content that the user provides, and assist the user with a variety of tasks using natural language."
-           "The visual content will be provided with the following format: <Image>visual content</Image>.",
-    roles=("Human", "Assistant"),
-    messages=(
-    ),
-    offset=0,
-    sep_style=SeparatorStyle.SINGLE,
-    sep="###",
-    version="v0_mmtag",
-)
-
-conv_llava_v1 = Conversation(
-    system="A chat between a curious human and an artificial intelligence assistant. "
-           "The assistant gives helpful, detailed, and polite answers to the human's questions.",
-    roles=("USER", "ASSISTANT"),
-    version="v1",
-    messages=(),
-    offset=0,
-    sep_style=SeparatorStyle.TWO,
-    sep=" ",
-    sep2="</s>",
-)
-
-conv_llava_v1_mmtag = Conversation(
-    system="A chat between a curious user and an artificial intelligence assistant. "
-           "The assistant is able to understand the visual content that the user provides, and assist the user with a variety of tasks using natural language."
-           "The visual content will be provided with the following format: <Image>visual content</Image>.",
-    roles=("USER", "ASSISTANT"),
-    messages=(),
-    offset=0,
-    sep_style=SeparatorStyle.TWO,
-    sep=" ",
-    sep2="</s>",
-    version="v1_mmtag",
-)
-
-conv_mistral_instruct = Conversation(
-    system="",
-    roles=("USER", "ASSISTANT"),
-    version="llama_v2",
-    messages=(),
-    offset=0,
-    sep_style=SeparatorStyle.LLAMA_2,
-    sep="",
-    sep2="</s>",
-)
-
-conv_chatml_direct = Conversation(
-    system="""<|im_start|>system
-Answer the questions.""",
-    roles=("<|im_start|>user\n", "<|im_start|>assistant\n"),
-    version="mpt",
-    messages=(),
-    offset=0,
-    sep_style=SeparatorStyle.MPT,
-    sep="<|im_end|>",
-)
-
-default_conversation = conv_vicuna_v1
-conv_templates = {
-    "default": conv_vicuna_v0,
-    "v0": conv_vicuna_v0,
-    "v1": conv_vicuna_v1,
-    "vicuna_v1": conv_vicuna_v1,
-    "llama_2": conv_llama_2,
-    "mistral_instruct": conv_mistral_instruct,
-    "chatml_direct": conv_chatml_direct,
-    "mistral_direct": conv_chatml_direct,
-
-    "plain": conv_llava_plain,
-    "v0_plain": conv_llava_plain,
-    "llava_v0": conv_llava_v0,
-    "v0_mmtag": conv_llava_v0_mmtag,
-    "llava_v1": conv_llava_v1,
-    "v1_mmtag": conv_llava_v1_mmtag,
-    "llava_llama_2": conv_llava_llama_2,
-
-    "mpt": conv_mpt,
-}
-
-
-if __name__ == "__main__":
-    print(default_conversation.get_prompt())

dialoggen/llava/mm_utils.py

@@ -1,247 +0,0 @@
-from PIL import Image
-from io import BytesIO
-import base64
-import torch
-import math
-import ast
-
-from transformers import StoppingCriteria
-from llava.constants import IMAGE_TOKEN_INDEX
-
-
-def select_best_resolution(original_size, possible_resolutions):
-    """
-    Selects the best resolution from a list of possible resolutions based on the original size.
-
-    Args:
-        original_size (tuple): The original size of the image in the format (width, height).
-        possible_resolutions (list): A list of possible resolutions in the format [(width1, height1), (width2, height2), ...].
-
-    Returns:
-        tuple: The best fit resolution in the format (width, height).
-    """
-    original_width, original_height = original_size
-    best_fit = None
-    max_effective_resolution = 0
-    min_wasted_resolution = float('inf')
-
-    for width, height in possible_resolutions:
-        scale = min(width / original_width, height / original_height)
-        downscaled_width, downscaled_height = int(original_width * scale), int(original_height * scale)
-        effective_resolution = min(downscaled_width * downscaled_height, original_width * original_height)
-        wasted_resolution = (width * height) - effective_resolution
-
-        if effective_resolution > max_effective_resolution or (effective_resolution == max_effective_resolution and wasted_resolution < min_wasted_resolution):
-            max_effective_resolution = effective_resolution
-            min_wasted_resolution = wasted_resolution
-            best_fit = (width, height)
-
-    return best_fit
-
-
-def resize_and_pad_image(image, target_resolution):
-    """
-    Resize and pad an image to a target resolution while maintaining aspect ratio.
-
-    Args:
-        image (PIL.Image.Image): The input image.
-        target_resolution (tuple): The target resolution (width, height) of the image.
-
-    Returns:
-        PIL.Image.Image: The resized and padded image.
-    """
-    original_width, original_height = image.size
-    target_width, target_height = target_resolution
-
-    scale_w = target_width / original_width
-    scale_h = target_height / original_height
-
-    if scale_w < scale_h:
-        new_width = target_width
-        new_height = min(math.ceil(original_height * scale_w), target_height)
-    else:
-        new_height = target_height
-        new_width = min(math.ceil(original_width * scale_h), target_width)
-
-    # Resize the image
-    resized_image = image.resize((new_width, new_height))
-
-    new_image = Image.new('RGB', (target_width, target_height), (0, 0, 0))
-    paste_x = (target_width - new_width) // 2
-    paste_y = (target_height - new_height) // 2
-    new_image.paste(resized_image, (paste_x, paste_y))
-
-    return new_image
-
-
-def divide_to_patches(image, patch_size):
-    """
-    Divides an image into patches of a specified size.
-
-    Args:
-        image (PIL.Image.Image): The input image.
-        patch_size (int): The size of each patch.
-
-    Returns:
-        list: A list of PIL.Image.Image objects representing the patches.
-    """
-    patches = []
-    width, height = image.size
-    for i in range(0, height, patch_size):
-        for j in range(0, width, patch_size):
-            box = (j, i, j + patch_size, i + patch_size)
-            patch = image.crop(box)
-            patches.append(patch)
-
-    return patches
-
-
-def get_anyres_image_grid_shape(image_size, grid_pinpoints, patch_size):
-    """
-    Calculate the shape of the image patch grid after the preprocessing for images of any resolution.
-
-    Args:
-        image_size (tuple): The size of the input image in the format (width, height).
-        grid_pinpoints (str): A string representation of a list of possible resolutions.
-        patch_size (int): The size of each image patch.
-
-    Returns:
-        tuple: The shape of the image patch grid in the format (width, height).
-    """
-    if type(grid_pinpoints) is list:
-        possible_resolutions = grid_pinpoints
-    else:
-        possible_resolutions = ast.literal_eval(grid_pinpoints)
-    width, height = select_best_resolution(image_size, possible_resolutions)
-    return width // patch_size, height // patch_size
-
-
-def process_anyres_image(image, processor, grid_pinpoints):
-    """
-    Process an image with variable resolutions.
-
-    Args:
-        image (PIL.Image.Image): The input image to be processed.
-        processor: The image processor object.
-        grid_pinpoints (str): A string representation of a list of possible resolutions.
-
-    Returns:
-        torch.Tensor: A tensor containing the processed image patches.
-    """
-    if type(grid_pinpoints) is list:
-        possible_resolutions = grid_pinpoints
-    else:
-        possible_resolutions = ast.literal_eval(grid_pinpoints)
-    best_resolution = select_best_resolution(image.size, possible_resolutions)
-    image_padded = resize_and_pad_image(image, best_resolution)
-
-    patches = divide_to_patches(image_padded, processor.crop_size['height'])
-
-    image_original_resize = image.resize((processor.size['shortest_edge'], processor.size['shortest_edge']))
-
-    image_patches = [image_original_resize] + patches
-    image_patches = [processor.preprocess(image_patch, return_tensors='pt')['pixel_values'][0]
-                     for image_patch in image_patches]
-    return torch.stack(image_patches, dim=0)
-
-
-def load_image_from_base64(image):
-    return Image.open(BytesIO(base64.b64decode(image)))
-
-
-def expand2square(pil_img, background_color):
-    width, height = pil_img.size
-    if width == height:
-        return pil_img
-    elif width > height:
-        result = Image.new(pil_img.mode, (width, width), background_color)
-        result.paste(pil_img, (0, (width - height) // 2))
-        return result
-    else:
-        result = Image.new(pil_img.mode, (height, height), background_color)
-        result.paste(pil_img, ((height - width) // 2, 0))
-        return result
-
-
-def process_images(images, image_processor, model_cfg):
-    image_aspect_ratio = getattr(model_cfg, "image_aspect_ratio", None)
-    new_images = []
-    if image_aspect_ratio == 'pad':
-        for image in images:
-            image = expand2square(image, tuple(int(x*255) for x in image_processor.image_mean))
-            image = image_processor.preprocess(image, return_tensors='pt')['pixel_values'][0]
-            new_images.append(image)
-    elif image_aspect_ratio == "anyres":
-        for image in images:
-            image = process_anyres_image(image, image_processor, model_cfg.image_grid_pinpoints)
-            new_images.append(image)
-    else:
-        return image_processor(images, return_tensors='pt')['pixel_values']
-    if all(x.shape == new_images[0].shape for x in new_images):
-        new_images = torch.stack(new_images, dim=0)
-    return new_images
-
-
-def tokenizer_image_token(prompt, tokenizer, image_token_index=IMAGE_TOKEN_INDEX, return_tensors=None):
-    prompt_chunks = [tokenizer(chunk).input_ids for chunk in prompt.split('<image>')]
-
-    def insert_separator(X, sep):
-        return [ele for sublist in zip(X, [sep]*len(X)) for ele in sublist][:-1]
-
-    input_ids = []
-    offset = 0
-    if len(prompt_chunks) > 0 and len(prompt_chunks[0]) > 0 and prompt_chunks[0][0] == tokenizer.bos_token_id:
-        offset = 1
-        input_ids.append(prompt_chunks[0][0])
-
-    for x in insert_separator(prompt_chunks, [image_token_index] * (offset + 1)):
-        input_ids.extend(x[offset:])
-
-    if return_tensors is not None:
-        if return_tensors == 'pt':
-            return torch.tensor(input_ids, dtype=torch.long)
-        raise ValueError(f'Unsupported tensor type: {return_tensors}')
-    return input_ids
-
-
-def get_model_name_from_path(model_path):
-    model_path = model_path.strip("/")
-    model_paths = model_path.split("/")
-    if model_paths[-1].startswith('checkpoint-'):
-        return model_paths[-2] + "_" + model_paths[-1]
-    else:
-        return model_paths[-1]
-
-class KeywordsStoppingCriteria(StoppingCriteria):
-    def __init__(self, keywords, tokenizer, input_ids):
-        self.keywords = keywords
-        self.keyword_ids = []
-        self.max_keyword_len = 0
-        for keyword in keywords:
-            cur_keyword_ids = tokenizer(keyword).input_ids
-            if len(cur_keyword_ids) > 1 and cur_keyword_ids[0] == tokenizer.bos_token_id:
-                cur_keyword_ids = cur_keyword_ids[1:]
-            if len(cur_keyword_ids) > self.max_keyword_len:
-                self.max_keyword_len = len(cur_keyword_ids)
-            self.keyword_ids.append(torch.tensor(cur_keyword_ids))
-        self.tokenizer = tokenizer
-        self.start_len = input_ids.shape[1]
-
-    def call_for_batch(self, output_ids: torch.LongTensor, scores: torch.FloatTensor, **kwargs) -> bool:
-        offset = min(output_ids.shape[1] - self.start_len, self.max_keyword_len)
-        self.keyword_ids = [keyword_id.to(output_ids.device) for keyword_id in self.keyword_ids]
-        for keyword_id in self.keyword_ids:
-            truncated_output_ids = output_ids[0, -keyword_id.shape[0]:]
-            if torch.equal(truncated_output_ids, keyword_id):
-                return True
-        outputs = self.tokenizer.batch_decode(output_ids[:, -offset:], skip_special_tokens=True)[0]
-        for keyword in self.keywords:
-            if keyword in outputs:
-                return True
-        return False
-
-    def __call__(self, output_ids: torch.LongTensor, scores: torch.FloatTensor, **kwargs) -> bool:
-        outputs = []
-        for i in range(output_ids.shape[0]):
-            outputs.append(self.call_for_batch(output_ids[i].unsqueeze(0), scores))
-        return all(outputs)

dialoggen/llava/model/__init__.py

@@ -1,6 +0,0 @@
-try:
-    from .language_model.llava_llama import LlavaLlamaForCausalLM, LlavaConfig
-    from .language_model.llava_mpt import LlavaMptForCausalLM, LlavaMptConfig
-    from .language_model.llava_mistral import LlavaMistralForCausalLM, LlavaMistralConfig
-except:
-    pass

dialoggen/llava/model/apply_delta.py

@@ -1,48 +0,0 @@
-"""
-Usage:
-python3 -m fastchat.model.apply_delta --base ~/model_weights/llama-7b --target ~/model_weights/vicuna-7b --delta lmsys/vicuna-7b-delta
-"""
-import argparse
-
-import torch
-from tqdm import tqdm
-from transformers import AutoTokenizer, AutoModelForCausalLM
-from llava import LlavaLlamaForCausalLM
-
-
-def apply_delta(base_model_path, target_model_path, delta_path):
-    print("Loading base model")
-    base = AutoModelForCausalLM.from_pretrained(
-        base_model_path, torch_dtype=torch.float16, low_cpu_mem_usage=True)
-
-    print("Loading delta")
-    delta = LlavaLlamaForCausalLM.from_pretrained(delta_path, torch_dtype=torch.float16, low_cpu_mem_usage=True)
-    delta_tokenizer = AutoTokenizer.from_pretrained(delta_path)
-
-    print("Applying delta")
-    for name, param in tqdm(delta.state_dict().items(), desc="Applying delta"):
-        if name not in base.state_dict():
-            assert name in ['model.mm_projector.weight', 'model.mm_projector.bias'], f'{name} not in base model'
-            continue
-        if param.data.shape == base.state_dict()[name].shape:
-            param.data += base.state_dict()[name]
-        else:
-            assert name in ['model.embed_tokens.weight', 'lm_head.weight'], \
-                f'{name} dimension mismatch: {param.data.shape} vs {base.state_dict()[name].shape}'
-            bparam = base.state_dict()[name]
-            param.data[:bparam.shape[0], :bparam.shape[1]] += bparam
-
-    print("Saving target model")
-    delta.save_pretrained(target_model_path)
-    delta_tokenizer.save_pretrained(target_model_path)
-
-
-if __name__ == "__main__":
-    parser = argparse.ArgumentParser()
-    parser.add_argument("--base-model-path", type=str, required=True)
-    parser.add_argument("--target-model-path", type=str, required=True)
-    parser.add_argument("--delta-path", type=str, required=True)
-
-    args = parser.parse_args()
-
-    apply_delta(args.base_model_path, args.target_model_path, args.delta_path)

dialoggen/llava/model/builder.py

@@ -1,166 +0,0 @@
-#    Copyright 2023 Haotian Liu
-#
-#    Licensed under the Apache License, Version 2.0 (the "License");
-#    you may not use this file except in compliance with the License.
-#    You may obtain a copy of the License at
-#
-#        http://www.apache.org/licenses/LICENSE-2.0
-#
-#    Unless required by applicable law or agreed to in writing, software
-#    distributed under the License is distributed on an "AS IS" BASIS,
-#    WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-#    See the License for the specific language governing permissions and
-#    limitations under the License.
-
-
-import os
-import warnings
-import shutil
-
-from transformers import AutoTokenizer, AutoModelForCausalLM, AutoConfig, BitsAndBytesConfig
-import torch
-from llava.model import *
-from llava.constants import DEFAULT_IMAGE_PATCH_TOKEN, DEFAULT_IM_START_TOKEN, DEFAULT_IM_END_TOKEN
-
-
-def load_pretrained_model(model_path, model_base, model_name, load_8bit=False, load_4bit=False, device_map="auto", device="cuda", use_flash_attn=False, llava_type_model=True, **kwargs):
-    kwargs = {"device_map": device_map, **kwargs}
-
-    if device != "cuda":
-        kwargs['device_map'] = {"": device}
-    if load_8bit:
-        kwargs['load_in_8bit'] = True
-    elif load_4bit:
-        kwargs['load_in_4bit'] = True
-        kwargs['quantization_config'] = BitsAndBytesConfig(
-            load_in_4bit=True,
-            bnb_4bit_compute_dtype=torch.float16,
-            bnb_4bit_use_double_quant=True,
-            bnb_4bit_quant_type='nf4'
-        )
-    else:
-        kwargs['torch_dtype'] = torch.float16
-
-    if use_flash_attn:
-        kwargs['attn_implementation'] = 'flash_attention_2'
-
-    if 'llava' in model_name.lower():
-        # Load LLaVA model
-        if 'lora' in model_name.lower() and model_base is None:
-            warnings.warn('There is `lora` in model name but no `model_base` is provided. If you are loading a LoRA model, please provide the `model_base` argument. Detailed instruction: https://github.com/haotian-liu/LLaVA#launch-a-model-worker-lora-weights-unmerged.')
-        if 'lora' in model_name.lower() and model_base is not None:
-            from llava.model.language_model.llava_llama import LlavaConfig
-            lora_cfg_pretrained = LlavaConfig.from_pretrained(model_path)
-            tokenizer = AutoTokenizer.from_pretrained(model_base, use_fast=False)
-            print('Loading LLaVA from base model...')
-            model = LlavaLlamaForCausalLM.from_pretrained(model_base, low_cpu_mem_usage=True, config=lora_cfg_pretrained, **kwargs)
-            token_num, tokem_dim = model.lm_head.out_features, model.lm_head.in_features
-            if model.lm_head.weight.shape[0] != token_num:
-                model.lm_head.weight = torch.nn.Parameter(torch.empty(token_num, tokem_dim, device=model.device, dtype=model.dtype))
-                model.model.embed_tokens.weight = torch.nn.Parameter(torch.empty(token_num, tokem_dim, device=model.device, dtype=model.dtype))
-
-            print('Loading additional LLaVA weights...')
-            if os.path.exists(os.path.join(model_path, 'non_lora_trainables.bin')):
-                non_lora_trainables = torch.load(os.path.join(model_path, 'non_lora_trainables.bin'), map_location='cpu')
-            else:
-                # this is probably from HF Hub
-                from huggingface_hub import hf_hub_download
-                def load_from_hf(repo_id, filename, subfolder=None):
-                    cache_file = hf_hub_download(
-                        repo_id=repo_id,
-                        filename=filename,
-                        subfolder=subfolder)
-                    return torch.load(cache_file, map_location='cpu')
-                non_lora_trainables = load_from_hf(model_path, 'non_lora_trainables.bin')
-            non_lora_trainables = {(k[11:] if k.startswith('base_model.') else k): v for k, v in non_lora_trainables.items()}
-            if any(k.startswith('model.model.') for k in non_lora_trainables):
-                non_lora_trainables = {(k[6:] if k.startswith('model.') else k): v for k, v in non_lora_trainables.items()}
-            model.load_state_dict(non_lora_trainables, strict=False)
-
-            from peft import PeftModel
-            print('Loading LoRA weights...')
-            model = PeftModel.from_pretrained(model, model_path)
-            print('Merging LoRA weights...')
-            model = model.merge_and_unload()
-            print('Model is loaded...')
-        elif model_base is not None:
-            # this may be mm projector only
-            print('Loading LLaVA from base model...')
-            if 'mpt' in model_name.lower():
-                if not os.path.isfile(os.path.join(model_path, 'configuration_mpt.py')):
-                    shutil.copyfile(os.path.join(model_base, 'configuration_mpt.py'), os.path.join(model_path, 'configuration_mpt.py'))
-                tokenizer = AutoTokenizer.from_pretrained(model_base, use_fast=True)
-                cfg_pretrained = AutoConfig.from_pretrained(model_path, trust_remote_code=True)
-                model = LlavaMptForCausalLM.from_pretrained(model_base, low_cpu_mem_usage=True, config=cfg_pretrained, **kwargs)
-            else:
-                tokenizer = AutoTokenizer.from_pretrained(model_base, use_fast=False)
-                cfg_pretrained = AutoConfig.from_pretrained(model_path)
-                model = LlavaLlamaForCausalLM.from_pretrained(model_base, low_cpu_mem_usage=True, config=cfg_pretrained, **kwargs)
-
-            mm_projector_weights = torch.load(os.path.join(model_path, 'mm_projector.bin'), map_location='cpu')
-            mm_projector_weights = {k: v.to(torch.float16) for k, v in mm_projector_weights.items()}
-            model.load_state_dict(mm_projector_weights, strict=False)
-        else:
-            if 'mpt' in model_name.lower():
-                tokenizer = AutoTokenizer.from_pretrained(model_path, use_fast=True)
-                model = LlavaMptForCausalLM.from_pretrained(model_path, low_cpu_mem_usage=True, **kwargs)
-            elif 'mistral' in model_name.lower():
-                tokenizer = AutoTokenizer.from_pretrained(model_path)
-                model = LlavaMistralForCausalLM.from_pretrained(
-                    model_path,
-                    low_cpu_mem_usage=True,
-                    **kwargs
-                )
-            else:
-                tokenizer = AutoTokenizer.from_pretrained(model_path, use_fast=False)
-                model = LlavaLlamaForCausalLM.from_pretrained(
-                    model_path,
-                    low_cpu_mem_usage=True,
-                    **kwargs
-                )
-    else:
-        # Load language model
-        if model_base is not None:
-            # PEFT model
-            from peft import PeftModel
-            tokenizer = AutoTokenizer.from_pretrained(model_base, use_fast=False)
-            model = AutoModelForCausalLM.from_pretrained(model_base, low_cpu_mem_usage=True, **kwargs)
-            print(f"Loading LoRA weights from {model_path}")
-            model = PeftModel.from_pretrained(model, model_path)
-            print(f"Merging weights")
-            model = model.merge_and_unload()
-            print('Convert to FP16...')
-            model.to(torch.float16)
-        else:
-            use_fast = False
-            if 'mpt' in model_name.lower():
-                tokenizer = AutoTokenizer.from_pretrained(model_path, use_fast=True)
-                model = AutoModelForCausalLM.from_pretrained(model_path, low_cpu_mem_usage=True, trust_remote_code=True, **kwargs)
-            else:
-                tokenizer = AutoTokenizer.from_pretrained(model_path, use_fast=False)
-                model = AutoModelForCausalLM.from_pretrained(model_path, low_cpu_mem_usage=True, **kwargs)
-
-    image_processor = None
-
-    if llava_type_model:
-        mm_use_im_start_end = getattr(model.config, "mm_use_im_start_end", False)
-        mm_use_im_patch_token = getattr(model.config, "mm_use_im_patch_token", True)
-        if mm_use_im_patch_token:
-            tokenizer.add_tokens([DEFAULT_IMAGE_PATCH_TOKEN], special_tokens=True)
-        if mm_use_im_start_end:
-            tokenizer.add_tokens([DEFAULT_IM_START_TOKEN, DEFAULT_IM_END_TOKEN], special_tokens=True)
-        model.resize_token_embeddings(len(tokenizer))
-
-        vision_tower = model.get_vision_tower()
-        if not vision_tower.is_loaded:
-            vision_tower.load_model(device_map=device_map)
-        if device_map != 'auto':
-            vision_tower.to(device=device_map, dtype=torch.float16)
-        image_processor = vision_tower.image_processor
-
-    if hasattr(model.config, "max_sequence_length"):
-        context_len = model.config.max_sequence_length
-    else:
-        context_len = 2048
-
-    return tokenizer, model, image_processor, context_len

dialoggen/llava/model/consolidate.py

@@ -1,29 +0,0 @@
-"""
-Usage:
-python3 -m llava.model.consolidate --src ~/model_weights/llava-7b --dst ~/model_weights/llava-7b_consolidate
-"""
-import argparse
-
-import torch
-from transformers import AutoTokenizer, AutoModelForCausalLM
-from llava.model import *
-from llava.model.utils import auto_upgrade
-
-
-def consolidate_ckpt(src_path, dst_path):
-    print("Loading model")
-    auto_upgrade(src_path)
-    src_model = AutoModelForCausalLM.from_pretrained(src_path, torch_dtype=torch.float16, low_cpu_mem_usage=True)
-    src_tokenizer = AutoTokenizer.from_pretrained(src_path, use_fast=False)
-    src_model.save_pretrained(dst_path)
-    src_tokenizer.save_pretrained(dst_path)
-
-
-if __name__ == "__main__":
-    parser = argparse.ArgumentParser()
-    parser.add_argument("--src", type=str, required=True)
-    parser.add_argument("--dst", type=str, required=True)
-
-    args = parser.parse_args()
-
-    consolidate_ckpt(args.src, args.dst)

dialoggen/llava/model/language_model/llava_llama.py

@@ -1,158 +0,0 @@
-#    Copyright 2023 Haotian Liu
-#
-#    Licensed under the Apache License, Version 2.0 (the "License");
-#    you may not use this file except in compliance with the License.
-#    You may obtain a copy of the License at
-#
-#        http://www.apache.org/licenses/LICENSE-2.0
-#
-#    Unless required by applicable law or agreed to in writing, software
-#    distributed under the License is distributed on an "AS IS" BASIS,
-#    WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-#    See the License for the specific language governing permissions and
-#    limitations under the License.
-
-
-from typing import List, Optional, Tuple, Union
-
-import torch
-import torch.nn as nn
-
-from transformers import AutoConfig, AutoModelForCausalLM, \
-                         LlamaConfig, LlamaModel, LlamaForCausalLM
-
-from transformers.modeling_outputs import CausalLMOutputWithPast
-from transformers.generation.utils import GenerateOutput
-
-from ..llava_arch import LlavaMetaModel, LlavaMetaForCausalLM
-
-
-class LlavaConfig(LlamaConfig):
-    model_type = "llava_llama"
-
-
-class LlavaLlamaModel(LlavaMetaModel, LlamaModel):
-    config_class = LlavaConfig
-
-    def __init__(self, config: LlamaConfig):
-        super(LlavaLlamaModel, self).__init__(config)
-
-
-class LlavaLlamaForCausalLM(LlamaForCausalLM, LlavaMetaForCausalLM):
-    config_class = LlavaConfig
-
-    def __init__(self, config):
-        super(LlamaForCausalLM, self).__init__(config)
-        self.model = LlavaLlamaModel(config)
-        self.pretraining_tp = config.pretraining_tp
-        self.vocab_size = config.vocab_size
-        self.lm_head = nn.Linear(config.hidden_size, config.vocab_size, bias=False)
-
-        # Initialize weights and apply final processing
-        self.post_init()
-
-    def get_model(self):
-        return self.model
-
-    def forward(
-        self,
-        input_ids: torch.LongTensor = None,
-        attention_mask: Optional[torch.Tensor] = None,
-        position_ids: Optional[torch.LongTensor] = None,
-        past_key_values: Optional[List[torch.FloatTensor]] = None,
-        inputs_embeds: Optional[torch.FloatTensor] = None,
-        labels: Optional[torch.LongTensor] = None,
-        use_cache: Optional[bool] = None,
-        output_attentions: Optional[bool] = None,
-        output_hidden_states: Optional[bool] = None,
-        images: Optional[torch.FloatTensor] = None,
-        image_sizes: Optional[List[List[int]]] = None,
-        return_dict: Optional[bool] = None,
-    ) -> Union[Tuple, CausalLMOutputWithPast]:
-
-        if inputs_embeds is None:
-            (
-                input_ids,
-                position_ids,
-                attention_mask,
-                past_key_values,
-                inputs_embeds,
-                labels
-            ) = self.prepare_inputs_labels_for_multimodal(
-                input_ids,
-                position_ids,
-                attention_mask,
-                past_key_values,
-                labels,
-                images,
-                image_sizes
-            )
-
-        return super().forward(
-            input_ids=input_ids,
-            attention_mask=attention_mask,
-            position_ids=position_ids,
-            past_key_values=past_key_values,
-            inputs_embeds=inputs_embeds,
-            labels=labels,
-            use_cache=use_cache,
-            output_attentions=output_attentions,
-            output_hidden_states=output_hidden_states,
-            return_dict=return_dict
-        )
-
-    @torch.no_grad()
-    def generate(
-        self,
-        inputs: Optional[torch.Tensor] = None,
-        images: Optional[torch.Tensor] = None,
-        image_sizes: Optional[torch.Tensor] = None,
-        **kwargs,
-    ) -> Union[GenerateOutput, torch.LongTensor]:
-        position_ids = kwargs.pop("position_ids", None)
-        attention_mask = kwargs.pop("attention_mask", None)
-        if "inputs_embeds" in kwargs:
-            raise NotImplementedError("`inputs_embeds` is not supported")
-
-        if images is not None:
-            (
-                inputs,
-                position_ids,
-                attention_mask,
-                _,
-                inputs_embeds,
-                _
-            ) = self.prepare_inputs_labels_for_multimodal(
-                inputs,
-                position_ids,
-                attention_mask,
-                None,
-                None,
-                images,
-                image_sizes=image_sizes
-            )
-        else:
-            inputs_embeds = self.get_model().embed_tokens(inputs)
-
-        return super().generate(
-            position_ids=position_ids,
-            attention_mask=attention_mask,
-            inputs_embeds=inputs_embeds,
-            **kwargs
-        )
-
-    def prepare_inputs_for_generation(self, input_ids, past_key_values=None,
-                                      inputs_embeds=None, **kwargs):
-        images = kwargs.pop("images", None)
-        image_sizes = kwargs.pop("image_sizes", None)
-        inputs = super().prepare_inputs_for_generation(
-            input_ids, past_key_values=past_key_values, inputs_embeds=inputs_embeds, **kwargs
-        )
-        if images is not None:
-            inputs['images'] = images
-        if image_sizes is not None:
-            inputs['image_sizes'] = image_sizes
-        return inputs
-
-AutoConfig.register("llava_llama", LlavaConfig)
-AutoModelForCausalLM.register(LlavaConfig, LlavaLlamaForCausalLM)

dialoggen/llava/model/language_model/llava_mistral.py

@@ -1,158 +0,0 @@
-#    Copyright 2023 Haotian Liu
-#
-#    Licensed under the Apache License, Version 2.0 (the "License");
-#    you may not use this file except in compliance with the License.
-#    You may obtain a copy of the License at
-#
-#        http://www.apache.org/licenses/LICENSE-2.0
-#
-#    Unless required by applicable law or agreed to in writing, software
-#    distributed under the License is distributed on an "AS IS" BASIS,
-#    WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-#    See the License for the specific language governing permissions and
-#    limitations under the License.
-
-
-from typing import List, Optional, Tuple, Union
-
-import torch
-import torch.nn as nn
-from torch.nn import CrossEntropyLoss
-
-from transformers import AutoConfig, AutoModelForCausalLM, \
-                         MistralConfig, MistralModel, MistralForCausalLM
-
-from transformers.modeling_outputs import CausalLMOutputWithPast
-from transformers.generation.utils import GenerateOutput
-
-from ..llava_arch import LlavaMetaModel, LlavaMetaForCausalLM
-
-
-class LlavaMistralConfig(MistralConfig):
-    model_type = "llava_mistral"
-
-
-class LlavaMistralModel(LlavaMetaModel, MistralModel):
-    config_class = LlavaMistralConfig
-
-    def __init__(self, config: MistralConfig):
-        super(LlavaMistralModel, self).__init__(config)
-
-
-class LlavaMistralForCausalLM(MistralForCausalLM, LlavaMetaForCausalLM):
-    config_class = LlavaMistralConfig
-
-    def __init__(self, config):
-        super(MistralForCausalLM, self).__init__(config)
-        self.model = LlavaMistralModel(config)
-
-        self.lm_head = nn.Linear(config.hidden_size, config.vocab_size, bias=False)
-
-        # Initialize weights and apply final processing
-        self.post_init()
-
-    def get_model(self):
-        return self.model
-
-    def forward(
-        self,
-        input_ids: torch.LongTensor = None,
-        attention_mask: Optional[torch.Tensor] = None,
-        position_ids: Optional[torch.LongTensor] = None,
-        past_key_values: Optional[List[torch.FloatTensor]] = None,
-        inputs_embeds: Optional[torch.FloatTensor] = None,
-        labels: Optional[torch.LongTensor] = None,
-        use_cache: Optional[bool] = None,
-        output_attentions: Optional[bool] = None,
-        output_hidden_states: Optional[bool] = None,
-        images: Optional[torch.FloatTensor] = None,
-        image_sizes: Optional[List[List[int]]] = None,
-        return_dict: Optional[bool] = None,
-    ) -> Union[Tuple, CausalLMOutputWithPast]:
-
-        if inputs_embeds is None:
-            (
-                input_ids,
-                position_ids,
-                attention_mask,
-                past_key_values,
-                inputs_embeds,
-                labels
-            ) = self.prepare_inputs_labels_for_multimodal(
-                input_ids,
-                position_ids,
-                attention_mask,
-                past_key_values,
-                labels,
-                images,
-                image_sizes
-            )
-
-        return super().forward(
-            input_ids=input_ids,
-            attention_mask=attention_mask,
-            position_ids=position_ids,
-            past_key_values=past_key_values,
-            inputs_embeds=inputs_embeds,
-            labels=labels,
-            use_cache=use_cache,
-            output_attentions=output_attentions,
-            output_hidden_states=output_hidden_states,
-            return_dict=return_dict
-        )
-
-    @torch.no_grad()
-    def generate(
-        self,
-        inputs: Optional[torch.Tensor] = None,
-        images: Optional[torch.Tensor] = None,
-        image_sizes: Optional[torch.Tensor] = None,
-        **kwargs,
-    ) -> Union[GenerateOutput, torch.LongTensor]:
-        position_ids = kwargs.pop("position_ids", None)
-        attention_mask = kwargs.pop("attention_mask", None)
-        if "inputs_embeds" in kwargs:
-            raise NotImplementedError("`inputs_embeds` is not supported")
-
-        if images is not None:
-            (
-                inputs,
-                position_ids,
-                attention_mask,
-                _,
-                inputs_embeds,
-                _
-            ) = self.prepare_inputs_labels_for_multimodal(
-                inputs,
-                position_ids,
-                attention_mask,
-                None,
-                None,
-                images,
-                image_sizes=image_sizes
-            )
-        else:
-            inputs_embeds = self.get_model().embed_tokens(inputs)
-
-        return super().generate(
-            position_ids=position_ids,
-            attention_mask=attention_mask,
-            inputs_embeds=inputs_embeds,
-            **kwargs
-        )
-
-    def prepare_inputs_for_generation(self, input_ids, past_key_values=None,
-                                      inputs_embeds=None, **kwargs):
-        images = kwargs.pop("images", None)
-        image_sizes = kwargs.pop("image_sizes", None)
-        inputs = super().prepare_inputs_for_generation(
-            input_ids, past_key_values=past_key_values, inputs_embeds=inputs_embeds, **kwargs
-        )
-        if images is not None:
-            inputs['images'] = images
-        if image_sizes is not None:
-            inputs['image_sizes'] = image_sizes
-        return inputs
-
-AutoConfig.register("llava_mistral", LlavaMistralConfig)
-AutoModelForCausalLM.register(LlavaMistralConfig, LlavaMistralForCausalLM)

dialoggen/llava/model/language_model/llava_mpt.py

@@ -1,97 +0,0 @@
-#    Copyright 2023 Haotian Liu
-#
-#    Licensed under the Apache License, Version 2.0 (the "License");
-#    you may not use this file except in compliance with the License.
-#    You may obtain a copy of the License at
-#
-#        http://www.apache.org/licenses/LICENSE-2.0
-#
-#    Unless required by applicable law or agreed to in writing, software
-#    distributed under the License is distributed on an "AS IS" BASIS,
-#    WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-#    See the License for the specific language governing permissions and
-#    limitations under the License.
-
-
-from typing import Optional, Tuple
-
-import torch
-
-from transformers import AutoConfig, AutoModelForCausalLM, \
-                         MptConfig, MptForCausalLM, MptModel
-from llava.model.llava_arch import LlavaMetaModel, LlavaMetaForCausalLM
-
-
-class LlavaMptConfig(MptConfig):
-    model_type = "llava_mpt"
-
-
-class LlavaMptModel(LlavaMetaModel, MptModel):
-    config_class = LlavaMptConfig
-
-    def __init__(self, config: MptConfig):
-        config.hidden_size = config.d_model
-        super(LlavaMptModel, self).__init__(config)
-    
-    def embed_tokens(self, x):
-        return self.wte(x)
-
-
-class LlavaMptForCausalLM(MptForCausalLM, LlavaMetaForCausalLM):
-    config_class = LlavaMptConfig
-    supports_gradient_checkpointing = True
-
-    def __init__(self, config):
-        super(MptForCausalLM, self).__init__(config)
-
-        self.transformer = LlavaMptModel(config)
-        self.lm_head = torch.nn.Linear(config.hidden_size, config.vocab_size, bias=False)
-
-        # Initialize weights and apply final processing
-        self.post_init()
-
-    def get_model(self):
-        return self.transformer
-
-    def _set_gradient_checkpointing(self, module, value=False):
-        if isinstance(module, LlavaMptModel):
-            module.gradient_checkpointing = value
-
-    def forward(
-        self,
-        input_ids: Optional[torch.LongTensor] = None,
-        past_key_values: Optional[Tuple[Tuple[torch.Tensor, torch.Tensor], ...]] = None,
-        attention_mask: Optional[torch.Tensor] = None,
-        inputs_embeds: Optional[torch.Tensor] = None,
-        labels: Optional[torch.Tensor] = None,
-        use_cache: Optional[bool] = None,
-        output_attentions: Optional[bool] = None,
-        output_hidden_states: Optional[bool] = None,
-        return_dict: Optional[bool] = None,
-        images=None):
-
-        input_ids, attention_mask, past_key_values, inputs_embeds, labels = self.prepare_inputs_labels_for_multimodal(input_ids, attention_mask, past_key_values, labels, images)
-        
-        return super().forward(
-            input_ids,
-            past_key_values=past_key_values,
-            attention_mask=attention_mask,
-            inputs_embeds=inputs_embeds,
-            labels=labels,
-            use_cache=use_cache,
-            output_attentions=output_attentions,
-            output_hidden_states=output_hidden_states,
-            return_dict=return_dict,
-        )
-
-    def prepare_inputs_for_generation(self, input_ids, past_key_values=None, inputs_embeds=None, **kwargs):
-        images = kwargs.pop("images", None)
-        _inputs = super().prepare_inputs_for_generation(
-            input_ids, past_key_values=past_key_values, inputs_embeds=inputs_embeds, **kwargs
-        )
-        _inputs['images'] = images
-        return _inputs
-
-
-AutoConfig.register("llava_mpt", LlavaMptConfig)
-AutoModelForCausalLM.register(LlavaMptConfig, LlavaMptForCausalLM)

dialoggen/llava/model/llava_arch.py

@@ -1,368 +0,0 @@
-#    Copyright 2023 Haotian Liu
-#
-#    Licensed under the Apache License, Version 2.0 (the "License");
-#    you may not use this file except in compliance with the License.
-#    You may obtain a copy of the License at
-#
-#        http://www.apache.org/licenses/LICENSE-2.0
-#
-#    Unless required by applicable law or agreed to in writing, software
-#    distributed under the License is distributed on an "AS IS" BASIS,
-#    WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-#    See the License for the specific language governing permissions and
-#    limitations under the License.
-
-
-from abc import ABC, abstractmethod
-
-import torch
-import torch.nn as nn
-
-from .multimodal_encoder.builder import build_vision_tower
-from .multimodal_projector.builder import build_vision_projector
-
-from llava.constants import IGNORE_INDEX, IMAGE_TOKEN_INDEX, DEFAULT_IMAGE_PATCH_TOKEN, DEFAULT_IM_START_TOKEN, DEFAULT_IM_END_TOKEN
-
-from llava.mm_utils import get_anyres_image_grid_shape
-
-
-class LlavaMetaModel:
-
-    def __init__(self, config):
-        super(LlavaMetaModel, self).__init__(config)
-
-        if hasattr(config, "mm_vision_tower"):
-            self.vision_tower = build_vision_tower(config, delay_load=True)
-            self.mm_projector = build_vision_projector(config)
-
-            if 'unpad' in getattr(config, 'mm_patch_merge_type', ''):
-                self.image_newline = nn.Parameter(
-                    torch.empty(config.hidden_size, dtype=self.dtype)
-                )
-
-    def get_vision_tower(self):
-        vision_tower = getattr(self, 'vision_tower', None)
-        if type(vision_tower) is list:
-            vision_tower = vision_tower[0]
-        return vision_tower
-
-    def initialize_vision_modules(self, model_args, fsdp=None):
-        vision_tower = model_args.vision_tower
-        mm_vision_select_layer = model_args.mm_vision_select_layer
-        mm_vision_select_feature = model_args.mm_vision_select_feature
-        pretrain_mm_mlp_adapter = model_args.pretrain_mm_mlp_adapter
-        mm_patch_merge_type = model_args.mm_patch_merge_type
-
-        self.config.mm_vision_tower = vision_tower
-
-        if self.get_vision_tower() is None:
-            vision_tower = build_vision_tower(model_args)
-
-            if fsdp is not None and len(fsdp) > 0:
-                self.vision_tower = [vision_tower]
-            else:
-                self.vision_tower = vision_tower
-        else:
-            if fsdp is not None and len(fsdp) > 0:
-                vision_tower = self.vision_tower[0]
-            else:
-                vision_tower = self.vision_tower
-            vision_tower.load_model()
-
-        self.config.use_mm_proj = True
-        self.config.mm_projector_type = getattr(model_args, 'mm_projector_type', 'linear')
-        self.config.mm_hidden_size = vision_tower.hidden_size
-        self.config.mm_vision_select_layer = mm_vision_select_layer
-        self.config.mm_vision_select_feature = mm_vision_select_feature
-        self.config.mm_patch_merge_type = mm_patch_merge_type
-
-        if getattr(self, 'mm_projector', None) is None:
-            self.mm_projector = build_vision_projector(self.config)
-
-            if 'unpad' in mm_patch_merge_type:
-                embed_std = 1 / torch.sqrt(torch.tensor(self.config.hidden_size, dtype=self.dtype))
-                self.image_newline = nn.Parameter(
-                    torch.randn(self.config.hidden_size, dtype=self.dtype) * embed_std
-                )
-        else:
-            # In case it is frozen by LoRA
-            for p in self.mm_projector.parameters():
-                p.requires_grad = True
-
-        if pretrain_mm_mlp_adapter is not None:
-            mm_projector_weights = torch.load(pretrain_mm_mlp_adapter, map_location='cpu')
-            def get_w(weights, keyword):
-                return {k.split(keyword + '.')[1]: v for k, v in weights.items() if keyword in k}
-
-            self.mm_projector.load_state_dict(get_w(mm_projector_weights, 'mm_projector'))
-
-
-def unpad_image(tensor, original_size):
-    """
-    Unpads a PyTorch tensor of a padded and resized image.
-
-    Args:
-    tensor (torch.Tensor): The image tensor, assumed to be in CxHxW format.
-    original_size (tuple): The original size of the image (height, width).
-
-    Returns:
-    torch.Tensor: The unpadded image tensor.
-    """
-    original_width, original_height = original_size
-    current_height, current_width = tensor.shape[1:]
-
-    original_aspect_ratio = original_width / original_height
-    current_aspect_ratio = current_width / current_height
-
-    if original_aspect_ratio > current_aspect_ratio:
-        scale_factor = current_width / original_width
-        new_height = int(original_height * scale_factor)
-        padding = (current_height - new_height) // 2
-        unpadded_tensor = tensor[:, padding:current_height - padding, :]
-    else:
-        scale_factor = current_height / original_height
-        new_width = int(original_width * scale_factor)
-        padding = (current_width - new_width) // 2
-        unpadded_tensor = tensor[:, :, padding:current_width - padding]
-
-    return unpadded_tensor
-
-
-class LlavaMetaForCausalLM(ABC):
-
-    @abstractmethod
-    def get_model(self):
-        pass
-
-    def get_vision_tower(self):
-        return self.get_model().get_vision_tower()
-
-    def encode_images(self, images):
-        image_features = self.get_model().get_vision_tower()(images)
-        image_features = self.get_model().mm_projector(image_features)
-        return image_features
-
-    def prepare_inputs_labels_for_multimodal(
-        self, input_ids, position_ids, attention_mask, past_key_values, labels,
-        images, image_sizes=None
-    ):
-        vision_tower = self.get_vision_tower()
-        if vision_tower is None or images is None or input_ids.shape[1] == 1:
-            return input_ids, position_ids, attention_mask, past_key_values, None, labels
-
-        if type(images) is list or images.ndim == 5:
-            if type(images) is list:
-                images = [x.unsqueeze(0) if x.ndim == 3 else x for x in images]
-            concat_images = torch.cat([image for image in images], dim=0)
-            image_features = self.encode_images(concat_images)
-            split_sizes = [image.shape[0] for image in images]
-            image_features = torch.split(image_features, split_sizes, dim=0)
-            mm_patch_merge_type = getattr(self.config, 'mm_patch_merge_type', 'flat')
-            image_aspect_ratio = getattr(self.config, 'image_aspect_ratio', 'square')
-            if mm_patch_merge_type == 'flat':
-                image_features = [x.flatten(0, 1) for x in image_features]
-            elif mm_patch_merge_type.startswith('spatial'):
-                new_image_features = []
-                for image_idx, image_feature in enumerate(image_features):
-                    if image_feature.shape[0] > 1:
-                        base_image_feature = image_feature[0]
-                        image_feature = image_feature[1:]
-                        height = width = self.get_vision_tower().num_patches_per_side
-                        assert height * width == base_image_feature.shape[0]
-                        if image_aspect_ratio == 'anyres':
-                            num_patch_width, num_patch_height = get_anyres_image_grid_shape(image_sizes[image_idx], self.config.image_grid_pinpoints, self.get_vision_tower().config.image_size)
-                            image_feature = image_feature.view(num_patch_height, num_patch_width, height, width, -1)
-                        else:
-                            raise NotImplementedError
-                        if 'unpad' in mm_patch_merge_type:
-                            image_feature = image_feature.permute(4, 0, 2, 1, 3).contiguous()
-                            image_feature = image_feature.flatten(1, 2).flatten(2, 3)
-                            image_feature = unpad_image(image_feature, image_sizes[image_idx])
-                            image_feature = torch.cat((
-                                image_feature,
-                                self.model.image_newline[:, None, None].expand(*image_feature.shape[:-1], 1).to(image_feature.device)
-                            ), dim=-1)
-                            image_feature = image_feature.flatten(1, 2).transpose(0, 1)
-                        else:
-                            image_feature = image_feature.permute(0, 2, 1, 3, 4).contiguous()
-                            image_feature = image_feature.flatten(0, 3)
-                        image_feature = torch.cat((base_image_feature, image_feature), dim=0)
-                    else:
-                        image_feature = image_feature[0]
-                        if 'unpad' in mm_patch_merge_type:
-                            image_feature = torch.cat((
-                                image_feature,
-                                self.model.image_newline[None].to(image_feature.device)
-                            ), dim=0)
-                    new_image_features.append(image_feature)
-                image_features = new_image_features
-            else:
-                raise ValueError(f"Unexpected mm_patch_merge_type: {self.config.mm_patch_merge_type}")
-        else:
-            image_features = self.encode_images(images)
-
-        # TODO: image start / end is not implemented here to support pretraining.
-        if getattr(self.config, 'tune_mm_mlp_adapter', False) and getattr(self.config, 'mm_use_im_start_end', False):
-            raise NotImplementedError
-
-        # Let's just add dummy tensors if they do not exist,
-        # it is a headache to deal with None all the time.
-        # But it is not ideal, and if you have a better idea,
-        # please open an issue / submit a PR, thanks.
-        _labels = labels
-        _position_ids = position_ids
-        _attention_mask = attention_mask
-        if attention_mask is None:
-            attention_mask = torch.ones_like(input_ids, dtype=torch.bool)
-        else:
-            attention_mask = attention_mask.bool()
-        if position_ids is None:
-            position_ids = torch.arange(0, input_ids.shape[1], dtype=torch.long, device=input_ids.device)
-        if labels is None:
-            labels = torch.full_like(input_ids, IGNORE_INDEX)
-
-        # remove the padding using attention_mask -- FIXME
-        _input_ids = input_ids
-        input_ids = [cur_input_ids[cur_attention_mask] for cur_input_ids, cur_attention_mask in zip(input_ids, attention_mask)]
-        labels = [cur_labels[cur_attention_mask] for cur_labels, cur_attention_mask in zip(labels, attention_mask)]
-
-        new_input_embeds = []
-        new_labels = []
-        cur_image_idx = 0
-        for batch_idx, cur_input_ids in enumerate(input_ids):
-            num_images = (cur_input_ids == IMAGE_TOKEN_INDEX).sum()
-            if num_images == 0:
-                cur_image_features = image_features[cur_image_idx]
-                cur_input_embeds_1 = self.get_model().embed_tokens(cur_input_ids)
-                cur_input_embeds = torch.cat([cur_input_embeds_1, cur_image_features[0:0]], dim=0)
-                new_input_embeds.append(cur_input_embeds)
-                new_labels.append(labels[batch_idx])
-                cur_image_idx += 1
-                continue
-
-            image_token_indices = [-1] + torch.where(cur_input_ids == IMAGE_TOKEN_INDEX)[0].tolist() + [cur_input_ids.shape[0]]
-            cur_input_ids_noim = []
-            cur_labels = labels[batch_idx]
-            cur_labels_noim = []
-            for i in range(len(image_token_indices) - 1):
-                cur_input_ids_noim.append(cur_input_ids[image_token_indices[i]+1:image_token_indices[i+1]])
-                cur_labels_noim.append(cur_labels[image_token_indices[i]+1:image_token_indices[i+1]])
-            split_sizes = [x.shape[0] for x in cur_labels_noim]
-            cur_input_embeds = self.get_model().embed_tokens(torch.cat(cur_input_ids_noim))
-            cur_input_embeds_no_im = torch.split(cur_input_embeds, split_sizes, dim=0)
-            cur_new_input_embeds = []
-            cur_new_labels = []
-
-            for i in range(num_images + 1):
-                cur_new_input_embeds.append(cur_input_embeds_no_im[i])
-                cur_new_labels.append(cur_labels_noim[i])
-                if i < num_images:
-                    cur_image_features = image_features[cur_image_idx]
-                    cur_image_idx += 1
-                    cur_new_input_embeds.append(cur_image_features)
-                    cur_new_labels.append(torch.full((cur_image_features.shape[0],), IGNORE_INDEX, device=cur_labels.device, dtype=cur_labels.dtype))
-
-            cur_new_input_embeds = [x.to(self.device) for x in cur_new_input_embeds]
-
-            cur_new_input_embeds = torch.cat(cur_new_input_embeds)
-            cur_new_labels = torch.cat(cur_new_labels)
-
-            new_input_embeds.append(cur_new_input_embeds)
-            new_labels.append(cur_new_labels)
-
-        # Truncate sequences to max length as image embeddings can make the sequence longer
-        tokenizer_model_max_length = getattr(self.config, 'tokenizer_model_max_length', None)
-        if tokenizer_model_max_length is not None:
-            new_input_embeds = [x[:tokenizer_model_max_length] for x in new_input_embeds]
-            new_labels = [x[:tokenizer_model_max_length] for x in new_labels]
-
-        # Combine them
-        max_len = max(x.shape[0] for x in new_input_embeds)
-        batch_size = len(new_input_embeds)
-
-        new_input_embeds_padded = []
-        new_labels_padded = torch.full((batch_size, max_len), IGNORE_INDEX, dtype=new_labels[0].dtype, device=new_labels[0].device)
-        attention_mask = torch.zeros((batch_size, max_len), dtype=attention_mask.dtype, device=attention_mask.device)
-        position_ids = torch.zeros((batch_size, max_len), dtype=position_ids.dtype, device=position_ids.device)
-
-        for i, (cur_new_embed, cur_new_labels) in enumerate(zip(new_input_embeds, new_labels)):
-            cur_len = cur_new_embed.shape[0]
-            if getattr(self.config, 'tokenizer_padding_side', 'right') == "left":
-                new_input_embeds_padded.append(torch.cat((
-                    torch.zeros((max_len - cur_len, cur_new_embed.shape[1]), dtype=cur_new_embed.dtype, device=cur_new_embed.device),
-                    cur_new_embed
-                ), dim=0))
-                if cur_len > 0:
-                    new_labels_padded[i, -cur_len:] = cur_new_labels
-                    attention_mask[i, -cur_len:] = True
-                    position_ids[i, -cur_len:] = torch.arange(0, cur_len, dtype=position_ids.dtype, device=position_ids.device)
-            else:
-                new_input_embeds_padded.append(torch.cat((
-                    cur_new_embed,
-                    torch.zeros((max_len - cur_len, cur_new_embed.shape[1]), dtype=cur_new_embed.dtype, device=cur_new_embed.device)
-                ), dim=0))
-                if cur_len > 0:
-                    new_labels_padded[i, :cur_len] = cur_new_labels
-                    attention_mask[i, :cur_len] = True
-                    position_ids[i, :cur_len] = torch.arange(0, cur_len, dtype=position_ids.dtype, device=position_ids.device)
-
-        new_input_embeds = torch.stack(new_input_embeds_padded, dim=0)
-
-        if _labels is None:
-            new_labels = None
-        else:
-            new_labels = new_labels_padded
-
-        if _attention_mask is None:
-            attention_mask = None
-        else:
-            attention_mask = attention_mask.to(dtype=_attention_mask.dtype)
-
-        if _position_ids is None:
-            position_ids = None
-
-        return None, position_ids, attention_mask, past_key_values, new_input_embeds, new_labels
-
-    def initialize_vision_tokenizer(self, model_args, tokenizer):
-        if model_args.mm_use_im_patch_token:
-            tokenizer.add_tokens([DEFAULT_IMAGE_PATCH_TOKEN], special_tokens=True)
-            self.resize_token_embeddings(len(tokenizer))
-
-        if model_args.mm_use_im_start_end:
-            num_new_tokens = tokenizer.add_tokens([DEFAULT_IM_START_TOKEN, DEFAULT_IM_END_TOKEN], special_tokens=True)
-            self.resize_token_embeddings(len(tokenizer))
-
-            if num_new_tokens > 0:
-                input_embeddings = self.get_input_embeddings().weight.data
-                output_embeddings = self.get_output_embeddings().weight.data
-
-                input_embeddings_avg = input_embeddings[:-num_new_tokens].mean(
-                    dim=0, keepdim=True)
-                output_embeddings_avg = output_embeddings[:-num_new_tokens].mean(
-                    dim=0, keepdim=True)
-
-                input_embeddings[-num_new_tokens:] = input_embeddings_avg
-                output_embeddings[-num_new_tokens:] = output_embeddings_avg
-
-            if model_args.tune_mm_mlp_adapter:
-                for p in self.get_input_embeddings().parameters():
-                    p.requires_grad = True
-                for p in self.get_output_embeddings().parameters():
-                    p.requires_grad = False
-
-            if model_args.pretrain_mm_mlp_adapter:
-                mm_projector_weights = torch.load(model_args.pretrain_mm_mlp_adapter, map_location='cpu')
-                embed_tokens_weight = mm_projector_weights['model.embed_tokens.weight']
-                assert num_new_tokens == 2
-                if input_embeddings.shape == embed_tokens_weight.shape:
-                    input_embeddings[-num_new_tokens:] = embed_tokens_weight[-num_new_tokens:]
-                elif embed_tokens_weight.shape[0] == num_new_tokens:
-                    input_embeddings[-num_new_tokens:] = embed_tokens_weight
-                else:
-                    raise ValueError(f"Unexpected embed_tokens_weight shape. Pretrained: {embed_tokens_weight.shape}. Current: {input_embeddings.shape}. Numer of new tokens: {num_new_tokens}.")
-        elif model_args.mm_use_im_patch_token:
-            if model_args.tune_mm_mlp_adapter:
-                for p in self.get_input_embeddings().parameters():
-                    p.requires_grad = False
-                for p in self.get_output_embeddings().parameters():
-                    p.requires_grad = False

dialoggen/llava/model/make_delta.py

@@ -1,52 +0,0 @@
-"""
-Usage:
-python3 -m llava.model.make_delta --base ~/model_weights/llama-7b --target ~/model_weights/llava-7b --delta ~/model_weights/llava-7b-delta --hub-repo-id liuhaotian/llava-7b-delta
-"""
-import argparse
-
-import torch
-from tqdm import tqdm
-from transformers import AutoTokenizer, AutoModelForCausalLM
-from llava.model.utils import auto_upgrade
-
-
-def make_delta(base_model_path, target_model_path, delta_path, hub_repo_id):
-    print("Loading base model")
-    base = AutoModelForCausalLM.from_pretrained(
-        base_model_path, torch_dtype=torch.float16, low_cpu_mem_usage=True)
-
-    print("Loading target model")
-    auto_upgrade(target_model_path)
-    target = AutoModelForCausalLM.from_pretrained(target_model_path, torch_dtype=torch.float16, low_cpu_mem_usage=True)
-
-    print("Calculating delta")
-    for name, param in tqdm(target.state_dict().items(), desc="Calculating delta"):
-        if name not in base.state_dict():
-            assert name in ['model.mm_projector.weight', 'model.mm_projector.bias'], f'{name} not in base model'
-            continue
-        if param.data.shape == base.state_dict()[name].shape:
-            param.data -= base.state_dict()[name]
-        else:
-            assert name in ['model.embed_tokens.weight', 'lm_head.weight'], f'{name} dimension mismatch: {param.data.shape} vs {base.state_dict()[name].shape}'
-            bparam = base.state_dict()[name]
-            param.data[:bparam.shape[0], :bparam.shape[1]] -= bparam
-
-    print("Saving delta")
-    if hub_repo_id:
-        kwargs = {"push_to_hub": True, "repo_id": hub_repo_id}
-    else:
-        kwargs = {}
-    target.save_pretrained(delta_path, **kwargs)
-    target_tokenizer = AutoTokenizer.from_pretrained(target_model_path)
-    target_tokenizer.save_pretrained(delta_path, **kwargs)
-
-
-if __name__ == "__main__":
-    parser = argparse.ArgumentParser()
-    parser.add_argument("--base-model-path", type=str, required=True)
-    parser.add_argument("--target-model-path", type=str, required=True)
-    parser.add_argument("--delta-path", type=str, required=True)
-    parser.add_argument("--hub-repo-id", type=str, default=None)
-    args = parser.parse_args()
-
-    make_delta(args.base_model_path, args.target_model_path, args.delta_path, args.hub_repo_id)

dialoggen/llava/model/multimodal_encoder/builder.py

@@ -1,11 +0,0 @@
-import os
-from .clip_encoder import CLIPVisionTower
-
-
-def build_vision_tower(vision_tower_cfg, **kwargs):
-    vision_tower = getattr(vision_tower_cfg, 'mm_vision_tower', getattr(vision_tower_cfg, 'vision_tower', None))
-    is_absolute_path_exists = os.path.exists(vision_tower)
-    if is_absolute_path_exists or vision_tower.startswith("openai") or vision_tower.startswith("laion") or "ShareGPT4V" in vision_tower:
-        return CLIPVisionTower(vision_tower, args=vision_tower_cfg, **kwargs)
-
-    raise ValueError(f'Unknown vision tower: {vision_tower}')

dialoggen/llava/model/multimodal_encoder/clip_encoder.py

@@ -1,88 +0,0 @@
-import torch
-import torch.nn as nn
-
-from transformers import CLIPVisionModel, CLIPImageProcessor, CLIPVisionConfig
-
-
-class CLIPVisionTower(nn.Module):
-    def __init__(self, vision_tower, args,  delay_load=False):
-        super().__init__()
-
-        self.is_loaded = False
-
-        self.vision_tower_name = vision_tower
-        self.select_layer = args.mm_vision_select_layer
-        self.select_feature = getattr(args, 'mm_vision_select_feature', 'patch')
-
-        if not delay_load:
-            self.load_model()
-        elif getattr(args, 'unfreeze_mm_vision_tower', False):
-            self.load_model()
-        else:
-            self.cfg_only = CLIPVisionConfig.from_pretrained(self.vision_tower_name)
-
-    def load_model(self, device_map=None):
-        if self.is_loaded:
-            print('{} is already loaded, `load_model` called again, skipping.'.format(self.vision_tower_name))
-            return
-
-        self.image_processor = CLIPImageProcessor.from_pretrained(self.vision_tower_name)
-        self.vision_tower = CLIPVisionModel.from_pretrained(self.vision_tower_name, device_map=device_map)
-        self.vision_tower.requires_grad_(False)
-
-        self.is_loaded = True
-
-    def feature_select(self, image_forward_outs):
-        image_features = image_forward_outs.hidden_states[self.select_layer]
-        if self.select_feature == 'patch':
-            image_features = image_features[:, 1:]
-        elif self.select_feature == 'cls_patch':
-            image_features = image_features
-        else:
-            raise ValueError(f'Unexpected select feature: {self.select_feature}')
-        return image_features
-
-    @torch.no_grad()
-    def forward(self, images):
-        if type(images) is list:
-            image_features = []
-            for image in images:
-                image_forward_out = self.vision_tower(image.to(device=self.device, dtype=self.dtype).unsqueeze(0), output_hidden_states=True)
-                image_feature = self.feature_select(image_forward_out).to(image.dtype)
-                image_features.append(image_feature)
-        else:
-            image_forward_outs = self.vision_tower(images.to(device=self.device, dtype=self.dtype), output_hidden_states=True)
-            image_features = self.feature_select(image_forward_outs).to(images.dtype)
-
-        return image_features
-
-    @property
-    def dummy_feature(self):
-        return torch.zeros(1, self.hidden_size, device=self.device, dtype=self.dtype)
-
-    @property
-    def dtype(self):
-        return self.vision_tower.dtype
-
-    @property
-    def device(self):
-        return self.vision_tower.device
-
-    @property
-    def config(self):
-        if self.is_loaded:
-            return self.vision_tower.config
-        else:
-            return self.cfg_only
-
-    @property
-    def hidden_size(self):
-        return self.config.hidden_size
-
-    @property
-    def num_patches_per_side(self):
-        return self.config.image_size // self.config.patch_size
-
-    @property
-    def num_patches(self):
-        return (self.config.image_size // self.config.patch_size) ** 2

dialoggen/llava/model/multimodal_projector/builder.py

@@ -1,51 +0,0 @@
-import torch
-import torch.nn as nn
-import re
-
-
-class IdentityMap(nn.Module):
-    def __init__(self):
-        super().__init__()
-
-    def forward(self, x, *args, **kwargs):
-        return x
-
-    @property
-    def config(self):
-        return {"mm_projector_type": 'identity'}
-
-
-class SimpleResBlock(nn.Module):
-    def __init__(self, channels):
-        super().__init__()
-        self.pre_norm = nn.LayerNorm(channels)
-
-        self.proj = nn.Sequential(
-            nn.Linear(channels, channels),
-            nn.GELU(),
-            nn.Linear(channels, channels)
-        )
-    def forward(self, x):
-        x = self.pre_norm(x)
-        return x + self.proj(x)
-
-
-def build_vision_projector(config, delay_load=False, **kwargs):
-    projector_type = getattr(config, 'mm_projector_type', 'linear')
-
-    if projector_type == 'linear':
-        return nn.Linear(config.mm_hidden_size, config.hidden_size)
-
-    mlp_gelu_match = re.match(r'^mlp(\d+)x_gelu$', projector_type)
-    if mlp_gelu_match:
-        mlp_depth = int(mlp_gelu_match.group(1))
-        modules = [nn.Linear(config.mm_hidden_size, config.hidden_size)]
-        for _ in range(1, mlp_depth):
-            modules.append(nn.GELU())
-            modules.append(nn.Linear(config.hidden_size, config.hidden_size))
-        return nn.Sequential(*modules)
-
-    if projector_type == 'identity':
-        return IdentityMap()
-
-    raise ValueError(f'Unknown projector type: {projector_type}')

dialoggen/llava/model/utils.py

@@ -1,20 +0,0 @@
-from transformers import AutoConfig
-
-
-def auto_upgrade(config):
-    cfg = AutoConfig.from_pretrained(config)
-    if 'llava' in config and 'llava' not in cfg.model_type:
-        assert cfg.model_type == 'llama'
-        print("You are using newer LLaVA code base, while the checkpoint of v0 is from older code base.")
-        print("You must upgrade the checkpoint to the new code base (this can be done automatically).")
-        confirm = input("Please confirm that you want to upgrade the checkpoint. [Y/N]")
-        if confirm.lower() in ["y", "yes"]:
-            print("Upgrading checkpoint...")
-            assert len(cfg.architectures) == 1
-            setattr(cfg.__class__, "model_type", "llava")
-            cfg.architectures[0] = 'LlavaLlamaForCausalLM'
-            cfg.save_pretrained(config)
-            print("Checkpoint upgraded.")
-        else:
-            print("Checkpoint upgrade aborted.")
-            exit(1)

dialoggen/llava/utils.py

@@ -1,126 +0,0 @@
-import datetime
-import logging
-import logging.handlers
-import os
-import sys
-
-import requests
-
-from llava.constants import LOGDIR
-
-server_error_msg = "**NETWORK ERROR DUE TO HIGH TRAFFIC. PLEASE REGENERATE OR REFRESH THIS PAGE.**"
-moderation_msg = "YOUR INPUT VIOLATES OUR CONTENT MODERATION GUIDELINES. PLEASE TRY AGAIN."
-
-handler = None
-
-
-def build_logger(logger_name, logger_filename):
-    global handler
-
-    formatter = logging.Formatter(
-        fmt="%(asctime)s | %(levelname)s | %(name)s | %(message)s",
-        datefmt="%Y-%m-%d %H:%M:%S",
-    )
-
-    # Set the format of root handlers
-    if not logging.getLogger().handlers:
-        logging.basicConfig(level=logging.INFO)
-    logging.getLogger().handlers[0].setFormatter(formatter)
-
-    # Redirect stdout and stderr to loggers
-    stdout_logger = logging.getLogger("stdout")
-    stdout_logger.setLevel(logging.INFO)
-    sl = StreamToLogger(stdout_logger, logging.INFO)
-    sys.stdout = sl
-
-    stderr_logger = logging.getLogger("stderr")
-    stderr_logger.setLevel(logging.ERROR)
-    sl = StreamToLogger(stderr_logger, logging.ERROR)
-    sys.stderr = sl
-
-    # Get logger
-    logger = logging.getLogger(logger_name)
-    logger.setLevel(logging.INFO)
-
-    # Add a file handler for all loggers
-    if handler is None:
-        os.makedirs(LOGDIR, exist_ok=True)
-        filename = os.path.join(LOGDIR, logger_filename)
-        handler = logging.handlers.TimedRotatingFileHandler(
-            filename, when='D', utc=True, encoding='UTF-8')
-        handler.setFormatter(formatter)
-
-        for name, item in logging.root.manager.loggerDict.items():
-            if isinstance(item, logging.Logger):
-                item.addHandler(handler)
-
-    return logger
-
-
-class StreamToLogger(object):
-    """
-    Fake file-like stream object that redirects writes to a logger instance.
-    """
-    def __init__(self, logger, log_level=logging.INFO):
-        self.terminal = sys.stdout
-        self.logger = logger
-        self.log_level = log_level
-        self.linebuf = ''
-
-    def __getattr__(self, attr):
-        return getattr(self.terminal, attr)
-
-    def write(self, buf):
-        temp_linebuf = self.linebuf + buf
-        self.linebuf = ''
-        for line in temp_linebuf.splitlines(True):
-            # From the io.TextIOWrapper docs:
-            #   On output, if newline is None, any '\n' characters written
-            #   are translated to the system default line separator.
-            # By default sys.stdout.write() expects '\n' newlines and then
-            # translates them so this is still cross platform.
-            if line[-1] == '\n':
-                self.logger.log(self.log_level, line.rstrip())
-            else:
-                self.linebuf += line
-
-    def flush(self):
-        if self.linebuf != '':
-            self.logger.log(self.log_level, self.linebuf.rstrip())
-        self.linebuf = ''
-
-
-def disable_torch_init():
-    """
-    Disable the redundant torch default initialization to accelerate model creation.
-    """
-    import torch
-    setattr(torch.nn.Linear, "reset_parameters", lambda self: None)
-    setattr(torch.nn.LayerNorm, "reset_parameters", lambda self: None)
-
-
-def violates_moderation(text):
-    """
-    Check whether the text violates OpenAI moderation API.
-    """
-    url = "https://api.openai.com/v1/moderations"
-    headers = {"Content-Type": "application/json",
-               "Authorization": "Bearer " + os.environ["OPENAI_API_KEY"]}
-    text = text.replace("\n", "")
-    data = "{" + '"input": ' + f'"{text}"' + "}"
-    data = data.encode("utf-8")
-    try:
-        ret = requests.post(url, headers=headers, data=data, timeout=5)
-        flagged = ret.json()["results"][0]["flagged"]
-    except requests.exceptions.RequestException as e:
-        flagged = False
-    except KeyError as e:
-        flagged = False
-
-    return flagged
-
-
-def pretty_print_semaphore(semaphore):
-    if semaphore is None:
-        return "None"
-    return f"Semaphore(value={semaphore._value}, locked={semaphore.locked()})"