change module name to deepseek_vl2

README.md

@@ -102,8 +102,8 @@ pip install -e .
 import torch
 from transformers import AutoModelForCausalLM
 
-from deepseek_vl.models import DeepseekVLV2Processor, DeepseekVLV2ForCausalLM
+from deepseek_vl2.models import DeepseekVLV2Processor, DeepseekVLV2ForCausalLM
-from deepseek_vl.utils.io import load_pil_images
+from deepseek_vl2.utils.io import load_pil_images
 
 
 # specify the path to the model

deepseek_vl2/models/processing_deepseek_vl_v2.py

@@ -559,7 +559,7 @@ class DeepseekVLV2Processor(ProcessorMixin):
                 for j in range(0, best_width, self.image_size):
                     images_list.append(
                         self.image_transform(local_view.crop((j, i, j + self.image_size, i + self.image_size))))
-
+            
             """record height / width crop num"""
             num_width_tiles, num_height_tiles = best_width // self.image_size, best_height // self.image_size
             images_spatial_crop.append([num_width_tiles, num_height_tiles])

deepseek_vl2/serve/app_modules/overwrites.py

@@ -22,8 +22,8 @@ from __future__ import annotations
 import logging
 from typing import List, Tuple
 
-from deepseek_vl.serve.app_modules.presets import gr
+from deepseek_vl2.serve.app_modules.presets import gr
-from deepseek_vl.serve.app_modules.utils import convert_asis, convert_mdtext, detect_converted_mark
+from deepseek_vl2.serve.app_modules.utils import convert_asis, convert_mdtext, detect_converted_mark
 
 
 def compact_text_chunks(self, prompt, text_chunks: List[str]) -> List[str]:
@@ -58,8 +58,8 @@ def postprocess(
     return temp
 
 
-with open("deepseek_vl/serve/assets/custom.js", "r", encoding="utf-8") as f, open(
+with open("deepseek_vl2/serve/assets/custom.js", "r", encoding="utf-8") as f, open(
-    "deepseek_vl/serve/assets/Kelpy-Codos.js", "r", encoding="utf-8"
+    "deepseek_vl2/serve/assets/Kelpy-Codos.js", "r", encoding="utf-8"
 ) as f2:
     customJS = f.read()
     kelpyCodos = f2.read()

deepseek_vl2/serve/app_modules/utils.py

@@ -35,7 +35,7 @@ from pygments import highlight
 from pygments.formatters import HtmlFormatter
 from pygments.lexers import ClassNotFound, get_lexer_by_name, guess_lexer
 
-from deepseek_vl.serve.app_modules.presets import (
+from deepseek_vl2.serve.app_modules.presets import (
     ALREADY_CONVERTED_MARK,
     BOX2COLOR,
     MAX_IMAGE_SIZE,
@@ -50,9 +50,9 @@ def configure_logger():
     logger.setLevel(logging.DEBUG)
 
     timestr = time.strftime("%Y%m%d-%H%M%S")
-    os.makedirs("deepseek_vl/serve/logs", exist_ok=True)
+    os.makedirs("deepseek_vl2/serve/logs", exist_ok=True)
     file_handler = logging.FileHandler(
-        f"deepseek_vl/serve/logs/{timestr}_gradio_log.log"
+        f"deepseek_vl2/serve/logs/{timestr}_gradio_log.log"
     )
     console_handler = logging.StreamHandler()
 
@@ -301,7 +301,7 @@ def parse_ref_bbox(response, image):
             text_x = box[0]
             text_y = box[1] - 20
             text_color = box_color
-            font = ImageFont.truetype('./deepseek_vl/serve/assets/simsun.ttc', size=20)
+            font = ImageFont.truetype('./deepseek_vl2/serve/assets/simsun.ttc', size=20)
             draw.text((text_x, text_y), label, font=font, fill=text_color)
 
         return image

deepseek_vl2/serve/inference.py

@@ -30,8 +30,8 @@ from transformers import (
     TextIteratorStreamer,
 )
 
-from deepseek_vl.models import DeepseekVLV2Processor, DeepseekVLV2ForCausalLM
+from deepseek_vl2.models import DeepseekVLV2Processor, DeepseekVLV2ForCausalLM
-from deepseek_vl.models.conversation import Conversation
+from deepseek_vl2.models.conversation import Conversation
 
 
 def load_model(model_path, dtype=torch.bfloat16):

deepseek_vl2/utils/io.py

@@ -27,8 +27,8 @@ from transformers import AutoModelForCausalLM
 
 def load_pretrained_model(model_path: str):
 
-    from deepseek_vl.models.processing_deepseek_vl_v2 import DeepseekVLV2Processor
+    from deepseek_vl2.models.processing_deepseek_vl_v2 import DeepseekVLV2Processor
-    from deepseek_vl.models.modeling_deepseek_vl_v2 import DeepseekVLV2ForCausalLM
+    from deepseek_vl2.models.modeling_deepseek_vl_v2 import DeepseekVLV2ForCausalLM
 
     vl_chat_processor = DeepseekVLV2Processor.from_pretrained(model_path)
     tokenizer = vl_chat_processor.tokenizer

inference.py

@@ -24,8 +24,8 @@ from transformers import AutoModelForCausalLM
 
 import PIL.Image
 
-from deepseek_vl.models import DeepseekVLV2ForCausalLM, DeepseekVLV2Processor
+from deepseek_vl2.models import DeepseekVLV2ForCausalLM, DeepseekVLV2Processor
-from deepseek_vl.serve.app_modules.utils import parse_ref_bbox
+from deepseek_vl2.serve.app_modules.utils import parse_ref_bbox
 
 
 def load_pil_images(conversations: List[Dict[str, str]]) -> List[PIL.Image.Image]:

pyproject.toml

@@ -3,12 +3,12 @@ requires = ["setuptools>=40.6.0", "wheel"]
 build-backend = "setuptools.build_meta"
 
 [project]
-name = "deepseek_vl"
+name = "deepseek_vl2"
 version = "1.0.0"
-description = "DeepSeek-VL"
+description = "DeepSeek-VL2"
 authors = [{name = "DeepSeek-AI"}]
 license = {file = "LICENSE-CODE"}
-urls = {homepage = "https://github.com/deepseek-ai/DeepSeek-VL"}
+urls = {homepage = "https://github.com/deepseek-ai/DeepSeek-VL2"}
 readme = "README.md"
 requires-python = ">=3.8"
 dependencies = [