Upload 3 files

config.json

@@ -1,11 +1,15 @@
 {
-  "_name_or_path": "/data/hypertext/xpkong/detvary/checkpoints/318-train_chartqax10-min_max_normalize_l1/checkpoint-51500/",
+  "_name_or_path": "kppkkp/OneChart",
   "_remove_final_layer_norm": false,
   "activation_dropout": 0.0,
   "activation_function": "relu",
   "architectures": [
-    "MMGPTOPTForCausalLM"
+    "OneChartOPTForCausalLM"
   ],
+  "auto_map": {
+    "AutoConfig": "modeling_OneChart.OneChartConfig",
+    "AutoModel": "modeling_OneChart.OneChartOPTForCausalLM"
+  },
   "attention_dropout": 0.0,
   "bos_token_id": 2,
   "do_layer_norm_before": true,
@@ -18,15 +22,15 @@
   "im_end_token": 50267,
   "im_patch_token": 50265,
   "im_start_token": 50266,
+  "number_token": 50268,
   "image_token_len": 256,
   "init_std": 0.02,
   "layer_norm_elementwise_affine": true,
   "layerdrop": 0.0,
   "max_position_embeddings": 4096,
-  "model_type": "mmgpt",
+  "model_type": "OneChart",
   "num_attention_heads": 12,
   "num_hidden_layers": 12,
-  "number_token": 50268,
   "pad_token_id": 1,
   "prefix": "</s>",
   "torch_dtype": "bfloat16",
@@ -34,7 +38,6 @@
   "use_cache": true,
   "use_im_start_end": true,
   "vision_select_layer": -2,
-  "vision_tower": "/mnt/host0/vit-large-patch14",
   "vocab_size": 50269,
   "word_embed_proj_dim": 768
-}
+}

modeling_OneChart.py

@@ -0,0 +1,496 @@
+from transformers import OPTConfig, OPTModel, OPTForCausalLM, StoppingCriteria, TextStreamer
+from transformers.modeling_outputs import BaseModelOutputWithPast, CausalLMOutputWithPast
+from typing import List, Optional, Tuple, Union
+import requests
+from PIL import Image
+from io import BytesIO
+import json
+import re
+import torch
+import numpy as np
+import torch.nn as nn
+from torch.nn import CrossEntropyLoss
+import torch.nn.functional as F
+from .sam_vision_b import build_SAM_vit_b
+from torchvision import transforms
+from torchvision.transforms.functional import InterpolationMode
+import dataclasses
+
+DEFAULT_IMAGE_TOKEN = "<image>"
+DEFAULT_IMAGE_PATCH_TOKEN = '<imgpad>'
+DEFAULT_IM_START_TOKEN = '<img>'
+DEFAULT_IM_END_TOKEN = '</img>'
+
+from enum import auto, Enum
+class SeparatorStyle(Enum):
+    """Different separator style."""
+    SINGLE = auto()
+    TWO = auto()
+    MPT = 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 = "<|im_end|>"
+    sep2: str = None
+    version: str = "Unknown"
+
+    skip_next: bool = False
+
+    def get_prompt(self):
+        if self.sep_style == SeparatorStyle.SINGLE:
+            ret = self.system + self.sep + '\n'
+            for role, message in self.messages:
+                if message:
+                    if type(message) is tuple:
+                        message, _, _ = message
+                    ret += role + ": " + message + self.sep
+                else:
+                    ret += role + ":"
+            return ret
+        elif self.sep_style == SeparatorStyle.TWO:
+            seps = [self.sep, self.sep2]
+            ret = self.system + seps[0]
+            for i, (role, message) in enumerate(self.messages):
+                if message:
+                    if type(message) is tuple:
+                        message, _, _ = message
+                    ret += role + ": " + message + seps[i % 2]
+                else:
+                    ret += role + ":"
+            return ret
+        if self.sep_style == SeparatorStyle.MPT:
+            if self.system:
+                ret = self.system + self.sep 
+            else:
+                ret = ''
+            for role, message in self.messages:
+                if message:
+                    if type(message) is tuple:
+                        message, _, _ = message
+                    ret += role + message + self.sep
+                else:
+                    ret += role
+            return ret
+        else:
+            raise ValueError(f"Invalid style: {self.sep_style}")
+
+
+    def append_message(self, role, message):
+        self.messages.append([role, message])
+
+    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)
+
+
+class KeywordsStoppingCriteria(StoppingCriteria):
+    def __init__(self, keywords, tokenizer, input_ids):
+        self.keywords = keywords
+        self.keyword_ids = [tokenizer(keyword).input_ids for keyword in keywords]
+        self.keyword_ids = [keyword_id[0] for keyword_id in self.keyword_ids if type(keyword_id) is list and len(keyword_id) == 1]
+        self.tokenizer = tokenizer
+        self.start_len = None
+        self.input_ids = input_ids
+
+    def __call__(self, output_ids: torch.LongTensor, scores: torch.FloatTensor, **kwargs) -> bool:
+        if self.start_len is None:
+            self.start_len = self.input_ids.shape[1]
+        else:
+            for keyword_id in self.keyword_ids:
+                if output_ids[0, -1] == keyword_id:
+                    return True
+            outputs = self.tokenizer.batch_decode(output_ids[:, self.start_len:], skip_special_tokens=True)[0]
+            for keyword in self.keywords:
+                if keyword in outputs:
+                    return True
+        return False
+    
+conv_vicuna_v1_1 = 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>",
+)
+
+class OneChartImageEvalProcessor:
+    def __init__(self, image_size=1024):
+        mean = (0., 0., 0.)
+        std = (1., 1., 1.)
+        self.normalize = transforms.Normalize(mean, std)
+        self.transform = transforms.Compose(
+            [
+                transforms.Resize(
+                    (image_size, image_size), interpolation=InterpolationMode.BICUBIC
+                ),
+                transforms.ToTensor(),
+                self.normalize,
+            ]
+        )
+    def __call__(self, item):
+        return self.transform(item)
+
+
+class OneChartConfig(OPTConfig):
+    model_type = "OneChart"
+
+class OneChartModel(OPTModel):
+    config_class = OneChartConfig
+
+    def __init__(self, config: OPTConfig):
+        super(OneChartModel, self).__init__(config)
+        self.vision_tower = build_SAM_vit_b()
+        self.mm_projector =  nn.Linear(1024, 768)
+
+    def embed_tokens(self, x):
+        return self.get_input_embeddings()(x)
+    
+    def forward(
+        self,
+        input_ids: torch.LongTensor = None,
+        attention_mask: Optional[torch.Tensor] = None,
+        past_key_values: Optional[List[torch.FloatTensor]] = None,
+        inputs_embeds: Optional[torch.FloatTensor] = None,
+        use_cache: Optional[bool] = None,
+        output_attentions: Optional[bool] = None,
+        output_hidden_states: Optional[bool] = None,
+        images: Optional[torch.FloatTensor] = None,
+        return_dict: Optional[bool] = None,
+    ) -> Union[Tuple, BaseModelOutputWithPast]:
+
+        if inputs_embeds is None:
+            inputs_embeds = self.embed_tokens(input_ids)
+
+        vision_tower_high = getattr(self, 'vision_tower', None)
+        if vision_tower_high is not None and (input_ids.shape[1] != 1 or self.training) and images is not None:
+            use_im_start_end = getattr(self.config, "use_im_start_end", -1)
+            vision_select_layer = getattr(self.config, "vision_select_layer", -1)
+            im_patch_token = getattr(self.config, "im_patch_token", -1)
+            im_start_token = getattr(self.config, "im_start_token", -1)
+            im_end_token = getattr(self.config, "im_end_token", -1)
+            freeze_vision_tower = getattr(self.config, "freeze_vision_tower", False)
+            
+            image_features = []
+            for image in images:
+                P, C, H, W = image.shape
+                if P == 1:
+                    with torch.set_grad_enabled(False):
+                        cnn_feature = vision_tower_high(image)
+                        cnn_feature = cnn_feature.flatten(2).permute(0, 2, 1) # 256*1024
+                    image_feature = self.mm_projector(cnn_feature)
+                    image_features.append(image_feature)
+                else:
+                    raise NotImplementedError("Batch inference needs to be implemented.")
+
+
+            use_im_start_end = True
+            new_input_embeds = []
+            for cur_input_ids, cur_input_embeds, cur_image_features in zip(input_ids, inputs_embeds, image_features):
+                if use_im_start_end:
+                    if (cur_input_ids == im_start_token).sum() != (cur_input_ids == im_end_token).sum():
+                        raise ValueError("The number of image start tokens and image end tokens should be the same.")
+                    
+                    image_start_tokens = torch.where(cur_input_ids == im_start_token)[0]
+                    for image_start_token_pos, per_cur_image_features in zip(image_start_tokens, cur_image_features):
+                        per_cur_image_features = per_cur_image_features.to(device=cur_input_embeds.device)
+                        num_patches = per_cur_image_features.shape[0]
+
+                        if cur_input_ids[image_start_token_pos + num_patches + 1] != im_end_token:
+                            raise ValueError("The image end token should follow the image start token.")
+                        
+                        cur_input_embeds = torch.cat(
+                            (
+                                cur_input_embeds[:image_start_token_pos+1], 
+                                per_cur_image_features, 
+                                cur_input_embeds[image_start_token_pos + num_patches + 1:]
+                            ), 
+                            dim=0
+                        )
+
+                    new_input_embeds.append(cur_input_embeds)
+                else:
+                    raise NotImplementedError
+
+            inputs_embeds = torch.stack(new_input_embeds, dim=0)
+
+        return super(OneChartModel, self).forward(
+            input_ids=None, attention_mask=attention_mask, past_key_values=past_key_values,
+            inputs_embeds=inputs_embeds, use_cache=use_cache,
+            output_attentions=output_attentions, output_hidden_states=output_hidden_states,
+            return_dict=return_dict
+        )
+
+
+class OneChartOPTForCausalLM(OPTForCausalLM):
+    config_class = OneChartConfig
+    def __init__(self, config):
+        super(OneChartOPTForCausalLM, self).__init__(config)
+        self.model = OneChartModel(config)
+        self.vocab_size = config.vocab_size
+        self.num_decoder = nn.Sequential(
+            nn.Linear(config.hidden_size, config.hidden_size // 2),
+            nn.ReLU(),
+            nn.Linear(config.hidden_size // 2, config.hidden_size // 2),
+            nn.ReLU(),
+            nn.Linear(config.hidden_size // 2, 256),
+        )
+        self.lm_head = nn.Linear(config.hidden_size, config.vocab_size, bias=False)
+        self.pred_locs = []
+        # Initialize weights and apply final processing
+        self.post_init()
+
+    def get_model(self):
+        return self.model
+
+    def forward(
+        self,
+        input_ids: Optional[torch.LongTensor] = None,
+        past_key_values: Optional[Tuple[Tuple[torch.Tensor]]] = None,
+        attention_mask: Optional[torch.FloatTensor] = None,
+        token_type_ids: Optional[torch.LongTensor] = None,
+        position_ids: Optional[torch.LongTensor] = None,
+        head_mask: Optional[torch.FloatTensor] = None,
+        inputs_embeds: Optional[torch.FloatTensor] = None,
+        encoder_hidden_states: Optional[torch.Tensor] = None,
+        encoder_attention_mask: 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,
+        return_dict: Optional[bool] = None,
+        loc_labels=None,
+    ) -> Union[Tuple, CausalLMOutputWithPast]:
+        output_attentions = output_attentions if output_attentions is not None else self.config.output_attentions
+        output_hidden_states = (
+            output_hidden_states if output_hidden_states is not None else self.config.output_hidden_states
+        )
+        return_dict = return_dict if return_dict is not None else self.config.use_return_dict
+
+        outputs  = self.model(
+            input_ids=input_ids,
+            past_key_values=past_key_values,
+            attention_mask=attention_mask,
+            inputs_embeds=inputs_embeds,
+            use_cache=use_cache,
+            output_attentions=output_attentions,
+            output_hidden_states=output_hidden_states,
+            images=images,
+            return_dict=return_dict
+        )
+
+        hidden_states = outputs[0]
+        if (loc_labels is not None) and len(loc_labels) > 0:
+            det_patch_token = torch.where(input_ids == self.config.number_token)[1][0]
+            pred_locs = self.num_decoder(hidden_states[:, det_patch_token, :]) # shape: [batch_size, 256]
+        
+        # inference时输出num_head预测的值
+        if not self.training:
+            try:
+                det_patch_token = torch.where(input_ids == self.config.number_token)[1][0]
+                pred_locs = self.num_decoder(hidden_states[:, det_patch_token, :]) # shape: [batch_size, 256]
+                self.pred_locs = pred_locs[0][:100].cpu().tolist()
+            except Exception as e:
+                pass
+
+        logits = self.lm_head(hidden_states)
+        logits = logits.float()
+
+        # logits
+        loss = None
+        if labels is not None:
+            # Shift so that tokens < n predict n
+            shift_logits = logits[..., :-1, :].contiguous()
+            shift_labels = labels[..., 1:].contiguous()
+            # Flatten the tokens
+            loss_fct = CrossEntropyLoss()
+            shift_logits = shift_logits.view(-1, self.config.vocab_size)
+            shift_labels = shift_labels.view(-1)
+            # Enable model parallelism
+            shift_labels = shift_labels.to(shift_logits.device)
+            loss = loss_fct(shift_logits, shift_labels)
+
+        if not return_dict:
+            output = (logits,) + outputs[1:]
+            return (loss,) + output if loss is not None else output
+
+        return CausalLMOutputWithPast(
+            loss=loss,
+            logits=logits,
+            past_key_values=outputs.past_key_values,
+            hidden_states=outputs.hidden_states,
+            attentions=outputs.attentions,
+        )
+
+    def prepare_inputs_for_generation(
+        self, input_ids, past_key_values=None, inputs_embeds=None, **kwargs
+    ):
+        token_type_ids = kwargs.get("token_type_ids", None)
+        if past_key_values:
+            input_ids = input_ids[:, -1].unsqueeze(-1)
+            if token_type_ids is not None:
+                token_type_ids = token_type_ids[:, -1].unsqueeze(-1)
+
+        attention_mask = kwargs.get("attention_mask", None)
+        position_ids = kwargs.get("position_ids", None)
+
+        if attention_mask is not None and position_ids is None:
+            position_ids = attention_mask.long().cumsum(-1) - 1
+            position_ids.masked_fill_(attention_mask == 0, 1)
+            if past_key_values:
+                position_ids = position_ids[:, -1].unsqueeze(-1)
+        else:
+            position_ids = None
+
+        if inputs_embeds is not None and past_key_values is None:
+            model_inputs = {"inputs_embeds": inputs_embeds}
+        else:
+            model_inputs = {"input_ids": input_ids}
+
+        model_inputs.update(
+            {
+                "past_key_values": past_key_values,
+                "use_cache": kwargs.get("use_cache"),
+                "position_ids": position_ids,
+                "attention_mask": attention_mask,
+                "token_type_ids": token_type_ids,
+                "images": kwargs.get("images", None),
+            }
+        )
+        return model_inputs
+
+
+    def load_image(self, 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 disable_torch_init(self):
+        """
+        Disable the redundant torch default initialization to accelerate model creation.
+        """
+        setattr(torch.nn.Linear, "reset_parameters", lambda self: None)
+        setattr(torch.nn.LayerNorm, "reset_parameters", lambda self: None)
+
+    def chat(self, tokenizer, image_file, reliable_check=True, print_prompt=False):
+        dtype=torch.bfloat16
+        device="cuda"
+        def list_json_value(json_dict):
+            rst_str = []
+            sort_flag = True
+            try:
+                for key, value in json_dict.items():
+                    if isinstance(value, dict):
+                        decimal_out = list_json_value(value)
+                        rst_str = rst_str + decimal_out
+                        sort_flag = False
+                    elif isinstance(value, list):
+                        return []
+                    else:
+                        if isinstance(value, float) or isinstance(value, int):
+                            rst_str.append(value)
+                        else:
+                            # num_value = value.replace("%", "").replace("$", "").replace(" ", "").replace(",", "")
+                            value = re.sub(r'\(\d+\)|\[\d+\]', '', value)
+                            num_value = re.sub(r'[^\d.-]', '', str(value)) 
+                            if num_value not in ["-", "*", "none", "None", ""]:
+                                rst_str.append(float(num_value))
+            except Exception as e:
+                print(f"Error: {e}")
+                # print(json_dict)
+                return []
+            # if len(rst_str) > 0:
+            #     rst_str = rst_str + [float(-1)]
+            return rst_str
+
+        def norm_(rst_list):
+            if len(rst_list) < 2:
+                return rst_list
+            min_vals = min(rst_list)
+            max_vals = max(rst_list)
+            rst_list = np.array(rst_list)
+            normalized_tensor = (rst_list - min_vals) / (max_vals - min_vals + 1e-9)
+            return list(normalized_tensor)
+        
+        self.disable_torch_init()
+        image_processor_high =  OneChartImageEvalProcessor(image_size=1024)
+        use_im_start_end = True
+        image_token_len = 256
+        image = self.load_image(image_file)
+        image_tensor_1 = image_processor_high(image).to(dtype=dtype, device=device)
+
+        query = 'Convert the key information of the chart to a python dict:'
+        qs = DEFAULT_IM_START_TOKEN + DEFAULT_IMAGE_PATCH_TOKEN * image_token_len + DEFAULT_IM_END_TOKEN + query + '\n'
+        conv = conv_vicuna_v1_1.copy()
+        conv.append_message(conv.roles[0], qs)
+        conv.append_message(conv.roles[1], None)
+        prompt = conv.get_prompt()
+
+        if print_prompt:
+            print(prompt)
+
+        inputs = tokenizer([prompt])
+        input_ids = torch.as_tensor(inputs.input_ids).to(device=device)
+        stop_str = '</s>'
+        keywords = [stop_str]
+        stopping_criteria = KeywordsStoppingCriteria(keywords, tokenizer, input_ids)
+        streamer = TextStreamer(tokenizer, skip_prompt=True, skip_special_tokens=True)
+
+        with torch.autocast(device, dtype=dtype):
+            output_ids = self.generate(
+                input_ids,
+                images=[image_tensor_1.unsqueeze(0).half()],
+                do_sample=False,
+                num_beams = 1,
+                # no_repeat_ngram_size = 20,
+                # streamer=streamer,
+                max_new_tokens=4096,
+                stopping_criteria=[stopping_criteria]
+                )
+        outputs = tokenizer.decode(output_ids[0, input_ids.shape[1]:], skip_special_tokens=True)
+        outputs = outputs.replace("<Number>", "")
+        outputs = outputs.strip()
+        if outputs.endswith(stop_str):
+            outputs = outputs[:-len(stop_str)]
+        response_str = outputs
+        
+        if reliable_check:
+            pred_nums = self.pred_locs
+            try:
+                outputs_json = json.loads(outputs)
+                list_v = list_json_value(outputs_json['values'])
+                list_v = [round(x,4) for x in norm_(list_v)]
+                gt_nums = torch.tensor(list_v).reshape(1,-1)
+                response_str = response_str + "\n<Chart>: " + str(pred_nums[:len(list_v)])
+                pred_nums_ = torch.tensor(pred_nums[:len(list_v)]).reshape(1,-1)
+                reliable_distence = F.l1_loss(pred_nums_, gt_nums)
+                response_str = response_str + "\nreliable_distence: " + str(reliable_distence)
+                if reliable_distence < 0.1:
+                    response_str = response_str + "\nAfter OneChart checking, this prediction is reliable."
+                else:
+                    response_str = response_str + "\nThis prediction may be has error! "
+            except Exception as e:
+                response_str = response_str + "\nThis prediction may be has error! "
+                response_str = response_str + "\n" + str(e)
+
+        return response_str

sam_vision_b.py

@@ -0,0 +1,468 @@
+import torch
+import torch.nn.functional as F
+from typing import Optional, Tuple, Type
+from functools import partial
+import torch.nn as nn
+from typing import Type
+
+
+
+class MLPBlock(nn.Module):
+    def __init__(
+        self,
+        embedding_dim: int,
+        mlp_dim: int,
+        act: Type[nn.Module] = nn.GELU,
+    ) -> None:
+        super().__init__()
+        self.lin1 = nn.Linear(embedding_dim, mlp_dim)
+        self.lin2 = nn.Linear(mlp_dim, embedding_dim)
+        self.act = act()
+
+    def forward(self, x: torch.Tensor) -> torch.Tensor:
+        return self.lin2(self.act(self.lin1(x)))
+
+
+
+class LayerNorm2d(nn.Module):
+    def __init__(self, num_channels: int, eps: float = 1e-6) -> None:
+        super().__init__()
+        self.weight = nn.Parameter(torch.ones(num_channels))
+        self.bias = nn.Parameter(torch.zeros(num_channels))
+        self.eps = eps
+
+    def forward(self, x: torch.Tensor) -> torch.Tensor:
+        u = x.mean(1, keepdim=True)
+        s = (x - u).pow(2).mean(1, keepdim=True)
+        x = (x - u) / torch.sqrt(s + self.eps)
+        x = self.weight[:, None, None] * x + self.bias[:, None, None]
+        return x
+
+
+
+class ImageEncoderViT(nn.Module):
+    def __init__(
+        self,
+        img_size: int = 1024,
+        patch_size: int = 16,
+        in_chans: int = 3,
+        embed_dim: int = 768,
+        depth: int = 12,
+        num_heads: int = 12,
+        mlp_ratio: float = 4.0,
+        out_chans: int = 256,
+        qkv_bias: bool = True,
+        norm_layer: Type[nn.Module] = nn.LayerNorm,
+        act_layer: Type[nn.Module] = nn.GELU,
+        use_abs_pos: bool = True,
+        use_rel_pos: bool = False,
+        rel_pos_zero_init: bool = True,
+        window_size: int = 0,
+        global_attn_indexes: Tuple[int, ...] = (),
+    ) -> None:
+        """
+        Args:
+            img_size (int): Input image size.
+            patch_size (int): Patch size.
+            in_chans (int): Number of input image channels.
+            embed_dim (int): Patch embedding dimension.
+            depth (int): Depth of ViT.
+            num_heads (int): Number of attention heads in each ViT block.
+            mlp_ratio (float): Ratio of mlp hidden dim to embedding dim.
+            qkv_bias (bool): If True, add a learnable bias to query, key, value.
+            norm_layer (nn.Module): Normalization layer.
+            act_layer (nn.Module): Activation layer.
+            use_abs_pos (bool): If True, use absolute positional embeddings.
+            use_rel_pos (bool): If True, add relative positional embeddings to the attention map.
+            rel_pos_zero_init (bool): If True, zero initialize relative positional parameters.
+            window_size (int): Window size for window attention blocks.
+            global_attn_indexes (list): Indexes for blocks using global attention.
+        """
+        super().__init__()
+        self.img_size = img_size
+
+        self.patch_embed = PatchEmbed(
+            kernel_size=(patch_size, patch_size),
+            stride=(patch_size, patch_size),
+            in_chans=in_chans,
+            embed_dim=embed_dim,
+        )
+
+        self.pos_embed: Optional[nn.Parameter] = None
+        if use_abs_pos:
+            # Initialize absolute positional embedding with pretrain image size.
+            self.pos_embed = nn.Parameter(
+                torch.zeros(1, img_size // patch_size, img_size // patch_size, embed_dim)
+            )
+
+        self.blocks = nn.ModuleList()
+        for i in range(depth):
+            block = Block(
+                dim=embed_dim,
+                num_heads=num_heads,
+                mlp_ratio=mlp_ratio,
+                qkv_bias=qkv_bias,
+                norm_layer=norm_layer,
+                act_layer=act_layer,
+                use_rel_pos=use_rel_pos,
+                rel_pos_zero_init=rel_pos_zero_init,
+                window_size=window_size if i not in global_attn_indexes else 0,
+                input_size=(img_size // patch_size, img_size // patch_size),
+            )
+            self.blocks.append(block)
+
+        self.neck = nn.Sequential(
+            nn.Conv2d(
+                embed_dim,
+                out_chans,
+                kernel_size=1,
+                bias=False,
+            ),
+            LayerNorm2d(out_chans),
+            nn.Conv2d(
+                out_chans,
+                out_chans,
+                kernel_size=3,
+                padding=1,
+                bias=False,
+            ),
+            LayerNorm2d(out_chans),
+        )
+
+        
+        self.net_2 = nn.Conv2d(256, 512, kernel_size=3, stride=2, padding=1, bias=False)
+        self.net_3 = nn.Conv2d(512, 1024, kernel_size=3, stride=2, padding=1, bias=False)
+
+    def forward(self, x: torch.Tensor) -> torch.Tensor:
+        x = self.patch_embed(x)
+        if self.pos_embed is not None:
+            x = x + self.pos_embed
+
+        for blk in self.blocks:
+            x = blk(x)
+
+        x = self.neck(x.permute(0, 3, 1, 2))
+        x = self.net_2(x)
+        x = self.net_3(x)
+
+
+        return x
+
+
+class Block(nn.Module):
+    """Transformer blocks with support of window attention and residual propagation blocks"""
+
+    def __init__(
+        self,
+        dim: int,
+        num_heads: int,
+        mlp_ratio: float = 4.0,
+        qkv_bias: bool = True,
+        norm_layer: Type[nn.Module] = nn.LayerNorm,
+        act_layer: Type[nn.Module] = nn.GELU,
+        use_rel_pos: bool = False,
+        rel_pos_zero_init: bool = True,
+        window_size: int = 0,
+        input_size: Optional[Tuple[int, int]] = None,
+    ) -> None:
+        """
+        Args:
+            dim (int): Number of input channels.
+            num_heads (int): Number of attention heads in each ViT block.
+            mlp_ratio (float): Ratio of mlp hidden dim to embedding dim.
+            qkv_bias (bool): If True, add a learnable bias to query, key, value.
+            norm_layer (nn.Module): Normalization layer.
+            act_layer (nn.Module): Activation layer.
+            use_rel_pos (bool): If True, add relative positional embeddings to the attention map.
+            rel_pos_zero_init (bool): If True, zero initialize relative positional parameters.
+            window_size (int): Window size for window attention blocks. If it equals 0, then
+                use global attention.
+            input_size (tuple(int, int) or None): Input resolution for calculating the relative
+                positional parameter size.
+        """
+        super().__init__()
+        self.norm1 = norm_layer(dim)
+        self.attn = Attention(
+            dim,
+            num_heads=num_heads,
+            qkv_bias=qkv_bias,
+            use_rel_pos=use_rel_pos,
+            rel_pos_zero_init=rel_pos_zero_init,
+            input_size=input_size if window_size == 0 else (window_size, window_size),
+        )
+
+        self.norm2 = norm_layer(dim)
+        self.mlp = MLPBlock(embedding_dim=dim, mlp_dim=int(dim * mlp_ratio), act=act_layer)
+
+        self.window_size = window_size
+
+    def forward(self, x: torch.Tensor) -> torch.Tensor:
+        shortcut = x
+        x = self.norm1(x)
+        # Window partition
+        if self.window_size > 0:
+            H, W = x.shape[1], x.shape[2]
+            x, pad_hw = window_partition(x, self.window_size)
+
+        x = self.attn(x)
+        # Reverse window partition
+        if self.window_size > 0:
+            x = window_unpartition(x, self.window_size, pad_hw, (H, W))
+
+        x = shortcut + x
+        x = x + self.mlp(self.norm2(x))
+
+        return x
+
+
+class Attention(nn.Module):
+    """Multi-head Attention block with relative position embeddings."""
+
+    def __init__(
+        self,
+        dim: int,
+        num_heads: int = 8,
+        qkv_bias: bool = True,
+        use_rel_pos: bool = False,
+        rel_pos_zero_init: bool = True,
+        input_size: Optional[Tuple[int, int]] = None,
+    ) -> None:
+        """
+        Args:
+            dim (int): Number of input channels.
+            num_heads (int): Number of attention heads.
+            qkv_bias (bool):  If True, add a learnable bias to query, key, value.
+            rel_pos (bool): If True, add relative positional embeddings to the attention map.
+            rel_pos_zero_init (bool): If True, zero initialize relative positional parameters.
+            input_size (tuple(int, int) or None): Input resolution for calculating the relative
+                positional parameter size.
+        """
+        super().__init__()
+        self.num_heads = num_heads
+        head_dim = dim // num_heads
+        self.scale = head_dim**-0.5
+
+        self.qkv = nn.Linear(dim, dim * 3, bias=qkv_bias)
+        self.proj = nn.Linear(dim, dim)
+
+        self.use_rel_pos = use_rel_pos
+        if self.use_rel_pos:
+            assert (
+                input_size is not None
+            ), "Input size must be provided if using relative positional encoding."
+            # initialize relative positional embeddings
+            self.rel_pos_h = nn.Parameter(torch.zeros(2 * input_size[0] - 1, head_dim))
+            self.rel_pos_w = nn.Parameter(torch.zeros(2 * input_size[1] - 1, head_dim))
+
+    def forward(self, x: torch.Tensor) -> torch.Tensor:
+        B, H, W, _ = x.shape
+        # qkv with shape (3, B, nHead, H * W, C)
+        qkv = self.qkv(x).reshape(B, H * W, 3, self.num_heads, -1).permute(2, 0, 3, 1, 4)
+        # q, k, v with shape (B * nHead, H * W, C)
+        q, k, v = qkv.reshape(3, B * self.num_heads, H * W, -1).unbind(0)
+
+        attn = (q * self.scale) @ k.transpose(-2, -1)
+
+        if self.use_rel_pos:
+            attn = add_decomposed_rel_pos(attn, q, self.rel_pos_h, self.rel_pos_w, (H, W), (H, W))
+
+        attn = attn.softmax(dim=-1)
+        x = (attn @ v).view(B, self.num_heads, H, W, -1).permute(0, 2, 3, 1, 4).reshape(B, H, W, -1)
+        x = self.proj(x)
+
+        return x
+
+
+def window_partition(x: torch.Tensor, window_size: int) -> Tuple[torch.Tensor, Tuple[int, int]]:
+    """
+    Partition into non-overlapping windows with padding if needed.
+    Args:
+        x (tensor): input tokens with [B, H, W, C].
+        window_size (int): window size.
+
+    Returns:
+        windows: windows after partition with [B * num_windows, window_size, window_size, C].
+        (Hp, Wp): padded height and width before partition
+    """
+    B, H, W, C = x.shape
+
+    pad_h = (window_size - H % window_size) % window_size
+    pad_w = (window_size - W % window_size) % window_size
+    if pad_h > 0 or pad_w > 0:
+        x = F.pad(x, (0, 0, 0, pad_w, 0, pad_h))
+    Hp, Wp = H + pad_h, W + pad_w
+
+    x = x.view(B, Hp // window_size, window_size, Wp // window_size, window_size, C)
+    windows = x.permute(0, 1, 3, 2, 4, 5).contiguous().view(-1, window_size, window_size, C)
+    return windows, (Hp, Wp)
+
+
+def window_unpartition(
+    windows: torch.Tensor, window_size: int, pad_hw: Tuple[int, int], hw: Tuple[int, int]
+) -> torch.Tensor:
+    """
+    Window unpartition into original sequences and removing padding.
+    Args:
+        windows (tensor): input tokens with [B * num_windows, window_size, window_size, C].
+        window_size (int): window size.
+        pad_hw (Tuple): padded height and width (Hp, Wp).
+        hw (Tuple): original height and width (H, W) before padding.
+
+    Returns:
+        x: unpartitioned sequences with [B, H, W, C].
+    """
+    Hp, Wp = pad_hw
+    H, W = hw
+    B = windows.shape[0] // (Hp * Wp // window_size // window_size)
+    x = windows.view(B, Hp // window_size, Wp // window_size, window_size, window_size, -1)
+    x = x.permute(0, 1, 3, 2, 4, 5).contiguous().view(B, Hp, Wp, -1)
+
+    if Hp > H or Wp > W:
+        x = x[:, :H, :W, :].contiguous()
+    return x
+
+
+def get_rel_pos(q_size: int, k_size: int, rel_pos: torch.Tensor) -> torch.Tensor:
+    """
+    Get relative positional embeddings according to the relative positions of
+        query and key sizes.
+    Args:
+        q_size (int): size of query q.
+        k_size (int): size of key k.
+        rel_pos (Tensor): relative position embeddings (L, C).
+
+    Returns:
+        Extracted positional embeddings according to relative positions.
+    """
+    max_rel_dist = int(2 * max(q_size, k_size) - 1)
+    # Interpolate rel pos if needed.
+    if rel_pos.shape[0] != max_rel_dist:
+        # Interpolate rel pos.
+        rel_pos_resized = F.interpolate(
+            rel_pos.reshape(1, rel_pos.shape[0], -1).permute(0, 2, 1),
+            size=max_rel_dist,
+            mode="linear",
+        )
+        rel_pos_resized = rel_pos_resized.reshape(-1, max_rel_dist).permute(1, 0)
+    else:
+        rel_pos_resized = rel_pos
+
+    # Scale the coords with short length if shapes for q and k are different.
+    q_coords = torch.arange(q_size)[:, None] * max(k_size / q_size, 1.0)
+    k_coords = torch.arange(k_size)[None, :] * max(q_size / k_size, 1.0)
+    relative_coords = (q_coords - k_coords) + (k_size - 1) * max(q_size / k_size, 1.0)
+
+    return rel_pos_resized[relative_coords.long()]
+
+
+def add_decomposed_rel_pos(
+    attn: torch.Tensor,
+    q: torch.Tensor,
+    rel_pos_h: torch.Tensor,
+    rel_pos_w: torch.Tensor,
+    q_size: Tuple[int, int],
+    k_size: Tuple[int, int],
+) -> torch.Tensor:
+    """
+    Args:
+        attn (Tensor): attention map.
+        q (Tensor): query q in the attention layer with shape (B, q_h * q_w, C).
+        rel_pos_h (Tensor): relative position embeddings (Lh, C) for height axis.
+        rel_pos_w (Tensor): relative position embeddings (Lw, C) for width axis.
+        q_size (Tuple): spatial sequence size of query q with (q_h, q_w).
+        k_size (Tuple): spatial sequence size of key k with (k_h, k_w).
+
+    Returns:
+        attn (Tensor): attention map with added relative positional embeddings.
+    """
+    q_h, q_w = q_size
+    k_h, k_w = k_size
+    Rh = get_rel_pos(q_h, k_h, rel_pos_h)
+    Rw = get_rel_pos(q_w, k_w, rel_pos_w)
+
+    B, _, dim = q.shape
+    r_q = q.reshape(B, q_h, q_w, dim)
+    rel_h = torch.einsum("bhwc,hkc->bhwk", r_q, Rh)
+    rel_w = torch.einsum("bhwc,wkc->bhwk", r_q, Rw)
+
+    attn = (
+        attn.view(B, q_h, q_w, k_h, k_w) + rel_h[:, :, :, :, None] + rel_w[:, :, :, None, :]
+    ).view(B, q_h * q_w, k_h * k_w)
+
+    return attn
+
+
+class PatchEmbed(nn.Module):
+    """
+    Image to Patch Embedding.
+    """
+
+    def __init__(
+        self,
+        kernel_size: Tuple[int, int] = (16, 16),
+        stride: Tuple[int, int] = (16, 16),
+        padding: Tuple[int, int] = (0, 0),
+        in_chans: int = 3,
+        embed_dim: int = 768,
+    ) -> None:
+        """
+        Args:
+            kernel_size (Tuple): kernel size of the projection layer.
+            stride (Tuple): stride of the projection layer.
+            padding (Tuple): padding size of the projection layer.
+            in_chans (int): Number of input image channels.
+            embed_dim (int): Patch embedding dimension.
+        """
+        super().__init__()
+
+        self.proj = nn.Conv2d(
+            in_chans, embed_dim, kernel_size=kernel_size, stride=stride, padding=padding
+        )
+
+    def forward(self, x: torch.Tensor) -> torch.Tensor:
+        x = self.proj(x)
+        # B C H W -> B H W C
+        x = x.permute(0, 2, 3, 1)
+        return x
+
+
+
+def build_SAM_vit_b(checkpoint=None):
+    return _build_SAM_vision(
+        encoder_embed_dim=768,
+        encoder_depth=12,
+        encoder_num_heads=12,
+        encoder_global_attn_indexes=[2, 5, 8, 11],
+        checkpoint=checkpoint,
+    )
+
+
+def _build_SAM_vision(
+    encoder_embed_dim,
+    encoder_depth,
+    encoder_num_heads,
+    encoder_global_attn_indexes,
+    checkpoint=None,
+):
+    prompt_embed_dim = 256
+    image_size = 1024
+    vit_patch_size = 16
+    image_embedding_size = image_size // vit_patch_size
+    image_encoder=ImageEncoderViT(
+            depth=encoder_depth,
+            embed_dim=encoder_embed_dim,
+            img_size=image_size,
+            mlp_ratio=4,
+            norm_layer=partial(torch.nn.LayerNorm, eps=1e-6),
+            num_heads=encoder_num_heads,
+            patch_size=vit_patch_size,
+            qkv_bias=True,
+            use_rel_pos=True,
+            global_attn_indexes=encoder_global_attn_indexes,
+            window_size=14,
+            out_chans=prompt_embed_dim,
+        )
+    
+
+    return image_encoder
+