Update README.md

README.md

@@ -21,10 +21,17 @@ We develop InternViT-6B-448px-V1-5 by continuing the pre-training of the strong
 Additionally, we enhance the data scale, quality, and diversity of the pre-training dataset, resulting in the powerful robustness, OCR capability, and high-resolution processing capability of our
 1.5 version model. 
 
-## Released Models
+## Model Details
+- **Model Type:** vision foundation model, feature backbone
+- **Model Stats:**
+  - Params (M): 5540 (the last 3 blocks are discarded)
+  - Image size: 448 x 448, training with 1 - 12 tiles
+- **Pretrain Dataset:** LAION-en, LAION-zh, COYO, GRIT, COCO, TextCaps, Objects365, OpenImages, All-Seeing, Wukong, LaionCOCO, CC3M, and OCR-related datasets. 
+To enhance the OCR capability of the model, we have incorporated additional OCR data alongside the general caption datasets. Specifically, we utilized PaddleOCR to perform Chinese OCR on images from Wukong and English OCR on images from LAION-COCO. 
+- **Note:** InternViT-6B originally had 48 blocks, and we found that using the output after the fourth-to-last block worked best for VLLM. For ease of use and to save GPU memory, we simply discarded the last 3 blocks. Now, the model has only 45 blocks and the number of parameters has been reduced from 5.9B to 5.5B. Therefore, if you want to build a VLLM based on this model, **please make use of the features from the last layer.**
 
+## Released Models
 ### Vision Foundation model
-
 | Model                   | Date       | Download                                                               | Note                             |
 | ----------------------- | ---------- | ---------------------------------------------------------------------- | -------------------------------- |
 | InternViT-6B-448px-V1.5 | 2024.04.20 | 🤗 [HF link](https://huggingface.co/OpenGVLab/InternViT-6B-448px-V1-5) | support dynamic resolution, super strong OCR (🔥new) |
@@ -34,7 +41,6 @@ Additionally, we enhance the data scale, quality, and diversity of the pre-train
 | InternVL-14B-224px      | 2023.12.22 | 🤗 [HF link](https://huggingface.co/OpenGVLab/InternVL-14B-224px)      | vision-language foundation model |
 
 ### Multimodal Large Language Model (MLLM)
-
 | Model                   | Date       | Download                                                                    | Note                               |
 | ----------------------- | ---------- | --------------------------------------------------------------------------- | ---------------------------------- |
 | InternVL-Chat-V1.5      | 2024.04.18 | 🤗 [HF link](https://huggingface.co/OpenGVLab/InternVL-Chat-V1-5)            | support 4K image; super strong OCR; Approaching the performance of GPT-4V and Gemini Pro on various benchmarks like MMMU, DocVQA, ChartQA, MathVista, etc. (🔥new)|
@@ -42,14 +48,6 @@ Additionally, we enhance the data scale, quality, and diversity of the pre-train
 | InternVL-Chat-V1.2      | 2024.02.11 | 🤗 [HF link](https://huggingface.co/OpenGVLab/InternVL-Chat-V1-2)            | scaling up LLM to 34B       |
 | InternVL-Chat-V1.1      | 2024.01.24 | 🤗 [HF link](https://huggingface.co/OpenGVLab/InternVL-Chat-V1-1)            | support Chinese and stronger OCR   |
 
-## Model Details
-- **Model Type:** vision foundation model, feature backbone
-- **Model Stats:**
-  - Params (M): 5540 (the last 3 blocks are discarded)
-  - Image size: 448 x 448, training with 1 - 12 tiles
-- **Pretrain Dataset:** LAION-en, LAION-zh, COYO, GRIT, COCO, TextCaps, Objects365, OpenImages, All-Seeing, Wukong, LaionCOCO, CC3M, and OCR-related datasets. 
-To enhance the OCR capability of the model, we have incorporated additional OCR data alongside the general caption datasets. Specifically, we utilized PaddleOCR to perform Chinese OCR on images from Wukong and English OCR on images from LAION-COCO. 
-- **Note:** InternViT-6B originally had 48 blocks, and we found that using the output after the fourth-to-last block worked best for VLLM. For ease of use and to save GPU memory, we simply discarded the last 3 blocks. Now, the model has only 45 blocks and the number of parameters has been reduced from 5.9B to 5.5B. Therefore, if you want to build a VLLM based on this model, **please make use of the features from the last layer.**
 ## Model Usage (Image Embeddings)
 
 ```python