zwgao commited on
Commit
37685b7
1 Parent(s): c5d3ba9

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +9 -11
README.md CHANGED
@@ -21,10 +21,17 @@ We develop InternViT-6B-448px-V1-5 by continuing the pre-training of the strong
21
  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
22
  1.5 version model.
23
 
24
- ## Released Models
 
 
 
 
 
 
 
25
 
 
26
  ### Vision Foundation model
27
-
28
  | Model | Date | Download | Note |
29
  | ----------------------- | ---------- | ---------------------------------------------------------------------- | -------------------------------- |
30
  | 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
34
  | InternVL-14B-224px | 2023.12.22 | 🤗 [HF link](https://huggingface.co/OpenGVLab/InternVL-14B-224px) | vision-language foundation model |
35
 
36
  ### Multimodal Large Language Model (MLLM)
37
-
38
  | Model | Date | Download | Note |
39
  | ----------------------- | ---------- | --------------------------------------------------------------------------- | ---------------------------------- |
40
  | 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
42
  | InternVL-Chat-V1.2 | 2024.02.11 | 🤗 [HF link](https://huggingface.co/OpenGVLab/InternVL-Chat-V1-2) | scaling up LLM to 34B |
43
  | InternVL-Chat-V1.1 | 2024.01.24 | 🤗 [HF link](https://huggingface.co/OpenGVLab/InternVL-Chat-V1-1) | support Chinese and stronger OCR |
44
 
45
- ## Model Details
46
- - **Model Type:** vision foundation model, feature backbone
47
- - **Model Stats:**
48
- - Params (M): 5540 (the last 3 blocks are discarded)
49
- - Image size: 448 x 448, training with 1 - 12 tiles
50
- - **Pretrain Dataset:** LAION-en, LAION-zh, COYO, GRIT, COCO, TextCaps, Objects365, OpenImages, All-Seeing, Wukong, LaionCOCO, CC3M, and OCR-related datasets.
51
- 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.
52
- - **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.**
53
  ## Model Usage (Image Embeddings)
54
 
55
  ```python
 
21
  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
22
  1.5 version model.
23
 
24
+ ## Model Details
25
+ - **Model Type:** vision foundation model, feature backbone
26
+ - **Model Stats:**
27
+ - Params (M): 5540 (the last 3 blocks are discarded)
28
+ - Image size: 448 x 448, training with 1 - 12 tiles
29
+ - **Pretrain Dataset:** LAION-en, LAION-zh, COYO, GRIT, COCO, TextCaps, Objects365, OpenImages, All-Seeing, Wukong, LaionCOCO, CC3M, and OCR-related datasets.
30
+ 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.
31
+ - **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.**
32
 
33
+ ## Released Models
34
  ### Vision Foundation model
 
35
  | Model | Date | Download | Note |
36
  | ----------------------- | ---------- | ---------------------------------------------------------------------- | -------------------------------- |
37
  | 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) |
 
41
  | InternVL-14B-224px | 2023.12.22 | 🤗 [HF link](https://huggingface.co/OpenGVLab/InternVL-14B-224px) | vision-language foundation model |
42
 
43
  ### Multimodal Large Language Model (MLLM)
 
44
  | Model | Date | Download | Note |
45
  | ----------------------- | ---------- | --------------------------------------------------------------------------- | ---------------------------------- |
46
  | 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)|
 
48
  | InternVL-Chat-V1.2 | 2024.02.11 | 🤗 [HF link](https://huggingface.co/OpenGVLab/InternVL-Chat-V1-2) | scaling up LLM to 34B |
49
  | InternVL-Chat-V1.1 | 2024.01.24 | 🤗 [HF link](https://huggingface.co/OpenGVLab/InternVL-Chat-V1-1) | support Chinese and stronger OCR |
50
 
 
 
 
 
 
 
 
 
51
  ## Model Usage (Image Embeddings)
52
 
53
  ```python