update

cli_chat.py

@@ -8,7 +8,8 @@ import readline
 from threading import Thread
 import torch
 from transformers import TextIteratorStreamer
-from deepseek_vlm.utils.io import load_pretrained_model
+
+from deepseek_vl.utils.io import load_pretrained_model
 
 
 def load_image(image_file):

deepseek_vl/models/clip_encoder.py

@@ -5,8 +5,8 @@ import torch.nn as nn
 import torchvision.transforms
 from einops import rearrange
 
-from deepseek_vlm.models.siglip_vit import create_siglip_vit
-from deepseek_vlm.models.sam import create_sam_vit
+from deepseek_vl.models.siglip_vit import create_siglip_vit
+from deepseek_vl.models.sam import create_sam_vit
 
 
 class CLIPVisionTower(nn.Module):

deepseek_vl/models/modeling_vlm.py

@@ -10,8 +10,8 @@ from transformers import (
     LlamaForCausalLM
 )
 
-from deepseek_vlm.models.projector import MlpProjector
-from deepseek_vlm.models.clip_encoder import CLIPVisionTower, HybridVisionTower
+from deepseek_vl.models.projector import MlpProjector
+from deepseek_vl.models.clip_encoder import CLIPVisionTower, HybridVisionTower
 
 
 def model_name_to_cls(cls_name):

deepseek_vl/models/processing_vlm.py

@@ -5,12 +5,11 @@ from PIL.Image import Image
 from typing import List, Dict, Union
 import torch
 
-from transformers import AutoTokenizer, AutoImageProcessor
 from transformers.processing_utils import ProcessorMixin
 from transformers import LlamaTokenizerFast
 
-from deepseek_vlm.models.image_processing_vlm import VLMImageProcessor
-from deepseek_vlm.utils.conversation import get_conv_template
+from deepseek_vl.models.image_processing_vlm import VLMImageProcessor
+from deepseek_vl.utils.conversation import get_conv_template
 
 
 class DictOutput(object):

deepseek_vl/utils/io.py

@@ -3,7 +3,8 @@ import PIL.Image
 from typing import Dict, List
 import torch
 from transformers import AutoModelForCausalLM
-from deepseek_vlm.models import VLChatProcessor, MultiModalityCausalLM
+
+from deepseek_vl.models import VLChatProcessor, MultiModalityCausalLM
 
 
 def load_pretrained_model(model_path: str):

inference.py

@@ -1,8 +1,8 @@
 import torch
 from transformers import AutoModelForCausalLM
 
-from deepseek_vlm.models import VLChatProcessor, MultiModalityCausalLM
-from deepseek_vlm.utils.io import load_pil_images
+from deepseek_vl.models import VLChatProcessor, MultiModalityCausalLM
+from deepseek_vl.utils.io import load_pil_images
 
 
 # specify the path to the model

setup.py

@@ -5,9 +5,9 @@ version = '1.0.0'
 print(version)
 
 setup(
-    name='deepseek_vlm',
+    name='deepseek_vl',
     version=version,
-    description='DeekSeel-VLM',
+    description='DeekSeel-VL',
     author='DeepSeek-AI',
     license='MIT',
     url='https://github.com/deepseek-ai/DeepSeek-VL',