osprey_arch.py

#    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 osprey.constants import IGNORE_INDEX, IMAGE_TOKEN_INDEX, DEFAULT_IMAGE_PATCH_TOKEN, DEFAULT_IM_START_TOKEN, DEFAULT_IM_END_TOKEN


class OspreyMetaModel:

    def __init__(self, config):
        super(OspreyMetaModel, self).__init__(config)

        if hasattr(config, "mm_vision_tower"):
            self.vision_tower = build_vision_tower(config, delay_load=False)
            self.mm_projector = build_vision_projector(config)

    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
        pretrain_mm_mlp_adapter = model_args.pretrain_mm_mlp_adapter

        if not hasattr(self.config, "mm_vision_tower"):
            self.config.mm_vision_tower = vision_tower

            vision_tower = build_vision_tower(model_args)

            if fsdp is not None and len(fsdp) > 0:
                self.vision_tower = [self.vision_tower]
            else:
                self.vision_tower = vision_tower

            self.config.use_mm_proj = True
            self.config.mm_projector_type = getattr(model_args, 'mm_projector_type', 'linear')
 
            self.mm_projector = build_vision_projector(self.config)

        if pretrain_mm_mlp_adapter is not None:
            print("***********load projector_weights********")
            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'))
        


class OspreyMetaForCausalLM(ABC):
    def __init__(self):
        super(OspreyMetaForCausalLM, self).__init__()

    @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, image_features_dict = self.get_model().get_vision_tower()(images)
        self.get_model().mm_projector.to(device=image_features.device, dtype=image_features.dtype)
        image_features = self.get_model().mm_projector(image_features)
        return image_features, image_features_dict

    def prepare_inputs_labels_for_multimodal(
        self, input_ids, masks, attention_mask, past_key_values, labels, images
    ):
        vision_tower = self.get_vision_tower()
        if vision_tower is None or images is None or input_ids.shape[1] == 1:
            if past_key_values is not None and vision_tower is not None and images is not None and input_ids.shape[1] == 1:
                attention_mask = torch.ones((attention_mask.shape[0], past_key_values[-1][-1].shape[-2] + 1), dtype=attention_mask.dtype, device=attention_mask.device)
            return input_ids, attention_mask, past_key_values, None, labels

        if type(images) is list or images.ndim == 5:
            concat_images = torch.cat([image for image in images], dim=0)
            image_features, image_features_dict = self.encode_images(concat_images)
            split_sizes = [image.shape[0] for image in images]
            image_features = torch.split(image_features, split_sizes, dim=0)
            image_features = [x.flatten(0, 1).to(concat_images.device) for x in image_features]
        else:
            image_features, image_features_dict = self.encode_images(images)
        

        mask_feats, pos_feats = self.mask_extractor(image_features_dict, masks)

        new_input_embeds = []
        new_labels = [] if labels is not None else None
        cur_image_idx = 0
        for batch_idx, cur_input_ids in enumerate(input_ids):
 
            if (cur_input_ids == IMAGE_TOKEN_INDEX).sum() == 0:
                # multimodal LLM, but the current sample is not multimodal
                # FIXME: this is a hacky fix, for deepspeed zero3 to work
                half_len = cur_input_ids.shape[0] // 2
                cur_image_features = image_features[cur_image_idx]
                cur_input_embeds_1 = self.get_model().embed_tokens(cur_input_ids[:half_len])
                cur_input_embeds_2 = self.get_model().embed_tokens(cur_input_ids[half_len:])
                cur_input_embeds = torch.cat([cur_input_embeds_1, cur_image_features[0:0], cur_input_embeds_2], dim=0)
                new_input_embeds.append(cur_input_embeds)
                if labels is not None:
                    new_labels.append(labels[batch_idx])
                cur_image_idx += 1
                continue
            image_token_indices = torch.where(cur_input_ids == IMAGE_TOKEN_INDEX)[0]
            cur_new_input_embeds = []
            if labels is not None:
                cur_labels = labels[batch_idx]
                cur_new_labels = []
                assert cur_labels.shape == cur_input_ids.shape
            while image_token_indices.numel() > 0:
                cur_image_features = image_features[cur_image_idx]
                image_token_start = image_token_indices[0]
                if getattr(self.config, 'tune_mm_mlp_adapter', False) and getattr(self.config, 'mm_use_im_start_end', False):
                    cur_new_input_embeds.append(self.get_model().embed_tokens(cur_input_ids[:image_token_start-1]).detach())
                    cur_new_input_embeds.append(self.get_model().embed_tokens(cur_input_ids[image_token_start-1:image_token_start]))
                    cur_new_input_embeds.append(cur_image_features)
                    cur_new_input_embeds.append(self.get_model().embed_tokens(cur_input_ids[image_token_start+1:image_token_start+2]))
                    if labels is not None:
                        cur_new_labels.append(cur_labels[:image_token_start])
                        cur_new_labels.append(torch.full((cur_image_features.shape[0],), IGNORE_INDEX, device=labels.device, dtype=labels.dtype))
                        cur_new_labels.append(cur_labels[image_token_start:image_token_start+1])
                        cur_labels = cur_labels[image_token_start+2:]
                else:
                    cur_new_input_embeds.append(self.get_model().embed_tokens(cur_input_ids[:image_token_start]))
                    cur_new_input_embeds.append(cur_image_features)
                    if labels is not None:
                        cur_new_labels.append(cur_labels[:image_token_start])
                        cur_new_labels.append(torch.full((cur_image_features.shape[0],), IGNORE_INDEX, device=labels.device, dtype=labels.dtype))
                        cur_labels = cur_labels[image_token_start+1:]
                cur_image_idx += 1
                if getattr(self.config, 'tune_mm_mlp_adapter', False) and getattr(self.config, 'mm_use_im_start_end', False):
                    cur_input_ids = cur_input_ids[image_token_start+2:]
                else:
                    cur_input_ids = cur_input_ids[image_token_start+1:]
                image_token_indices = torch.where(cur_input_ids == IMAGE_TOKEN_INDEX)[0]
            if cur_input_ids.numel() > 0:
                if getattr(self.config, 'tune_mm_mlp_adapter', False) and getattr(self.config, 'mm_use_im_start_end', False):
                    mask_idx = torch.nonzero(cur_input_ids==self.tokenizer.convert_tokens_to_ids(['<mask>'])[0])
    
                    _l = 0
                    for i, idx in enumerate(mask_idx):
                        cur_new_input_embeds.append(self.get_model().embed_tokens(cur_input_ids[_l:idx[0]]).detach())
                        ## mask
                        cur_new_input_embeds.append(mask_feats[batch_idx][i:i+1].detach())
                        ## pos
                        cur_new_input_embeds.append(pos_feats[batch_idx][i:i+1].detach())
                        if labels is not None:
                            cur_labels[idx[0]:idx[0]+2] = torch.full((2,), IGNORE_INDEX, device=labels.device, dtype=labels.dtype)
                        _l = idx[0]+2
                    if _l< len(cur_input_ids):
                        cur_new_input_embeds.append(self.get_model().embed_tokens(cur_input_ids[_l:]).detach())

                else:
                    
                    mask_idx = torch.nonzero(cur_input_ids==self.tokenizer.convert_tokens_to_ids(['<mask>'])[0])
                    assert len(mask_idx) == len(mask_feats[batch_idx]), "mask num not equal to mask feats"
                   
                    _l = 0
                    for i, idx in enumerate(mask_idx):
                        cur_raw_new_input_embeds = self.get_model().embed_tokens(cur_input_ids[_l:idx[0]])
                        cur_new_input_embeds.append(cur_raw_new_input_embeds)
                        ## mask
                        cur_new_input_embeds.append(mask_feats[batch_idx][i:i+1].to(cur_raw_new_input_embeds.dtype))
                        ## pos
                        cur_new_input_embeds.append(pos_feats[batch_idx][i:i+1].to(cur_raw_new_input_embeds.dtype))

                        if labels is not None:
                            cur_labels[idx[0]:idx[0]+2] = torch.full((2,), IGNORE_INDEX, device=labels.device, dtype=labels.dtype)

                        _l = idx[0]+2
                    if _l< len(cur_input_ids):
                        cur_new_input_embeds.append(self.get_model().embed_tokens(cur_input_ids[_l:]))

                if labels is not None:
                    cur_new_labels.append(cur_labels)
            cur_new_input_embeds = [x.to(device=self.device) for x in cur_new_input_embeds]
            cur_new_input_embeds = torch.cat(cur_new_input_embeds, dim=0)

            new_input_embeds.append(cur_new_input_embeds)
            if labels is not None:
                cur_new_labels = torch.cat(cur_new_labels, dim=0)
                new_labels.append(cur_new_labels)

        if any(x.shape != new_input_embeds[0].shape for x in new_input_embeds):
            max_len = max(x.shape[0] for x in new_input_embeds)

            new_input_embeds_align = []
            for cur_new_embed in new_input_embeds:
                cur_new_embed = torch.cat((cur_new_embed, torch.zeros((max_len - cur_new_embed.shape[0], cur_new_embed.shape[1]), dtype=cur_new_embed.dtype, device=cur_new_embed.device)), dim=0)
                new_input_embeds_align.append(cur_new_embed)
            new_input_embeds = torch.stack(new_input_embeds_align, dim=0)

            if labels is not None:
                new_labels_align = []
                _new_labels = new_labels
                for cur_new_label in new_labels:
                    cur_new_label = torch.cat((cur_new_label, torch.full((max_len - cur_new_label.shape[0],), IGNORE_INDEX, dtype=cur_new_label.dtype, device=cur_new_label.device)), dim=0)
                    new_labels_align.append(cur_new_label)
                new_labels = torch.stack(new_labels_align, dim=0)

            if attention_mask is not None:
                new_attention_mask = []
                for cur_attention_mask, cur_new_labels, cur_new_labels_align in zip(attention_mask, _new_labels, new_labels):
                    new_attn_mask_pad_left = torch.full((cur_new_labels.shape[0] - labels.shape[1],), True, dtype=attention_mask.dtype, device=attention_mask.device)
                    new_attn_mask_pad_right = torch.full((cur_new_labels_align.shape[0] - cur_new_labels.shape[0],), False, dtype=attention_mask.dtype, device=attention_mask.device)
                    cur_new_attention_mask = torch.cat((new_attn_mask_pad_left, cur_attention_mask, new_attn_mask_pad_right), dim=0)
                    new_attention_mask.append(cur_new_attention_mask)
                attention_mask = torch.stack(new_attention_mask, dim=0)
                assert attention_mask.shape == new_labels.shape
        else:
            new_input_embeds = torch.stack(new_input_embeds, dim=0)
            if labels is not None:
                new_labels  = torch.stack(new_labels, dim=0)

            if attention_mask is not None:
                new_attn_mask_pad_left = torch.full((attention_mask.shape[0], new_input_embeds.shape[1] - input_ids.shape[1]), True, dtype=attention_mask.dtype, device=attention_mask.device)
                attention_mask = torch.cat((new_attn_mask_pad_left, attention_mask), dim=1)
                assert attention_mask.shape == new_input_embeds.shape[:2]

        return None, 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))

        mask_tokens = ['<mask>', '<pos>']
        num_new_tokens = tokenizer.add_tokens(mask_tokens, special_tokens=True)

        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
        
        for m in self.modules():
            m.tokenizer = tokenizer