OpenGVLab
/

InternVL2_5-4B

@@ -81,7 +81,7 @@ The training pipeline for a single model in InternVL 2.5 is structured across th
 We introduce a progressive scaling strategy to align the vision encoder with LLMs efficiently. This approach trains with smaller LLMs first (e.g., 20B) to optimize foundational visual capabilities and cross-modal alignment before transferring the vision encoder to larger LLMs (e.g., 72B) without retraining. This reuse skips intermediate stages for larger models.
-![image/png](https://cdn-uploads.huggingface.co/production/uploads/64119264f0f81eb569e0d569/AVb_PSxhJq1z2eUFNYoqQ.png)
 Compared to Qwen2-VL's 1.4 trillion tokens, InternVL2.5-78B uses only 120 billion tokens—less than one-tenth. This strategy minimizes redundancy, maximizes pre-trained component reuse, and enables efficient training for complex vision-language tasks.
@@ -164,7 +164,7 @@ As shown in the following figure, from InternVL 1.5 to 2.0 and then to 2.5, the
 ### Video Understanding
-![image/png](https://cdn-uploads.huggingface.co/production/uploads/64119264f0f81eb569e0d569/uD5aYt2wNYL94Xn8MOVih.png)
 ## Evaluation on Language Capability
@@ -510,10 +510,10 @@ Many repositories now support fine-tuning of the InternVL series models, includi
 ### LMDeploy
-LMDeploy is a toolkit for compressing, deploying, and serving LLM, developed by the MMRazor and MMDeploy teams.
 ```sh
-pip install lmdeploy>=0.5.3
 ```
 LMDeploy abstracts the complex inference process of multi-modal Vision-Language Models (VLM) into an easy-to-use pipeline, similar to the Large Language Model (LLM) inference pipeline.
@@ -537,8 +537,6 @@ If `ImportError` occurs while executing this case, please install the required d
 When dealing with multiple images, you can put them all in one list. Keep in mind that multiple images will lead to a higher number of input tokens, and as a result, the size of the context window typically needs to be increased.
-question = 'Describe this video in detail.'
 ```python
 from lmdeploy import pipeline, TurbomindEngineConfig
 from lmdeploy.vl import load_image
@@ -602,7 +600,7 @@ print(sess.response.text)
 LMDeploy's `api_server` enables models to be easily packed into services with a single command. The provided RESTful APIs are compatible with OpenAI's interfaces. Below are an example of service startup:
 ```shell
-lmdeploy serve api_server OpenGVLab/InternVL2_5-4B --backend turbomind --server-port 23333
 ```
 To use the OpenAI-style interface, you need to install OpenAI:

 We introduce a progressive scaling strategy to align the vision encoder with LLMs efficiently. This approach trains with smaller LLMs first (e.g., 20B) to optimize foundational visual capabilities and cross-modal alignment before transferring the vision encoder to larger LLMs (e.g., 72B) without retraining. This reuse skips intermediate stages for larger models.
+![image/png](https://cdn-uploads.huggingface.co/production/uploads/64006c09330a45b03605bba3/UoNUyS7ctN5pBxNv9KnzH.png)
 Compared to Qwen2-VL's 1.4 trillion tokens, InternVL2.5-78B uses only 120 billion tokens—less than one-tenth. This strategy minimizes redundancy, maximizes pre-trained component reuse, and enables efficient training for complex vision-language tasks.
 ### Video Understanding
+![image/png](https://cdn-uploads.huggingface.co/production/uploads/64006c09330a45b03605bba3/tcwH-i1qc8H16En-7AZ5M.png)
 ## Evaluation on Language Capability
 ### LMDeploy
+LMDeploy is a toolkit for compressing, deploying, and serving LLMs & VLMs.
 ```sh
+pip install lmdeploy>=0.6.4
 ```
 LMDeploy abstracts the complex inference process of multi-modal Vision-Language Models (VLM) into an easy-to-use pipeline, similar to the Large Language Model (LLM) inference pipeline.
 When dealing with multiple images, you can put them all in one list. Keep in mind that multiple images will lead to a higher number of input tokens, and as a result, the size of the context window typically needs to be increased.
 ```python
 from lmdeploy import pipeline, TurbomindEngineConfig
 from lmdeploy.vl import load_image
 LMDeploy's `api_server` enables models to be easily packed into services with a single command. The provided RESTful APIs are compatible with OpenAI's interfaces. Below are an example of service startup:
 ```shell
+lmdeploy serve api_server OpenGVLab/InternVL2_5-4B --server-port 23333
 ```
 To use the OpenAI-style interface, you need to install OpenAI: