Yirany commited on
Commit
5c266fb
Β·
verified Β·
1 Parent(s): 10bde04

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +8 -4
README.md CHANGED
@@ -3,11 +3,11 @@ pipeline_tag: text-generation
3
  ---
4
 
5
  ## MiniCPM-V
6
- **MiniCPM-V** (i.e., OmniLMM-3B)is an efficient version with promising performance for deployment. The model is built based on [MiniCPM-2B](https://huggingface.co/openbmb/MiniCPM-2B-sft-bf16) and SigLip-400M, connected by a perceiver resampler. Notable features of MiniCPM-V include:
7
 
8
- - πŸš€ **High Efficiency.**
9
 
10
- MiniCPM-V can be **efficiently deployed on most GPU cards and personal computers**, and **even on edge devices such as mobile phones**. In terms of visual encoding, we compress the image representations into 64 tokens via a perceiver resampler, which is significantly fewer than other LMMs based on MLP architecture (typically > 512 tokens). This allows MiniCPM-V to operate with **much less memory cost and higher speed during inference**.
11
 
12
  - πŸ”₯ **Promising Performance.**
13
 
@@ -15,7 +15,7 @@ pipeline_tag: text-generation
15
 
16
  - πŸ™Œ **Bilingual Support.**
17
 
18
- MiniCPM-V is **the first edge-deployable LMM supporting bilingual multimodal interaction in English and Chinese**. This is achieved by generalizing multimodal capabilities across languages, a technique from our ICLR 2024 spotlight [paper](https://arxiv.org/abs/2308.12038).
19
 
20
  ### Evaluation
21
 
@@ -107,6 +107,9 @@ pipeline_tag: text-generation
107
  ## Demo
108
  Click here to try out the Demo of [MiniCPM-V](http://120.92.209.146:80).
109
 
 
 
 
110
 
111
  ## Usage
112
  Requirements: tested on python 3.10
@@ -146,6 +149,7 @@ print(res)
146
 
147
 
148
  ## License
 
149
  #### Model License
150
  * The code in this repo is released according to [Apache-2.0](https://github.com/OpenBMB/MiniCPM/blob/main/LICENSE)
151
  * The usage of MiniCPM-V's parameters is subject to ["General Model License Agreement - Source Notes - Publicity Restrictions - Commercial License"](https://github.com/OpenBMB/General-Model-License/blob/main/)
 
3
  ---
4
 
5
  ## MiniCPM-V
6
+ **MiniCPM-V** (i.e., OmniLMM-3B)is an efficient version with promising performance for deployment. The model is built based on SigLip-400M and [MiniCPM-2.4B](https://github.com/OpenBMB/MiniCPM/), connected by a perceiver resampler. Notable features of OmniLMM-3B include:
7
 
8
+ - ⚑️ **High Efficiency.**
9
 
10
+ MiniCPM-V can be **efficiently deployed on most GPU cards and personal computers**, and **even on end devices such as mobile phones**. In terms of visual encoding, we compress the image representations into 64 tokens via a perceiver resampler, which is significantly fewer than other LMMs based on MLP architecture (typically > 512 tokens). This allows OmniLMM-3B to operate with **much less memory cost and higher speed during inference**.
11
 
12
  - πŸ”₯ **Promising Performance.**
13
 
 
15
 
16
  - πŸ™Œ **Bilingual Support.**
17
 
18
+ MiniCPM-V is **the first end-deployable LMM supporting bilingual multimodal interaction in English and Chinese**. This is achieved by generalizing multimodal capabilities across languages, a technique from the ICLR 2024 spotlight [paper](https://arxiv.org/abs/2308.12038).
19
 
20
  ### Evaluation
21
 
 
107
  ## Demo
108
  Click here to try out the Demo of [MiniCPM-V](http://120.92.209.146:80).
109
 
110
+ ## Deployment on Mobile Phone
111
+ Currently MiniCPM-V (i.e., OmniLMM-3B) can be deployed on mobile phones with Android and Harmony operating systems. πŸš€ Try it out [here](https://github.com/OpenBMB/mlc-MiniCPM).
112
+
113
 
114
  ## Usage
115
  Requirements: tested on python 3.10
 
149
 
150
 
151
  ## License
152
+
153
  #### Model License
154
  * The code in this repo is released according to [Apache-2.0](https://github.com/OpenBMB/MiniCPM/blob/main/LICENSE)
155
  * The usage of MiniCPM-V's parameters is subject to ["General Model License Agreement - Source Notes - Publicity Restrictions - Commercial License"](https://github.com/OpenBMB/General-Model-License/blob/main/)