Update README.md

README.md

@@ -3,9 +3,11 @@ library_name: transformers
 license: bsd-3-clause
 base_model:
 - OpenGVLab/InternVL2_5-1B
+- OpenGVLab/InternVL2_5-1B-MPO
 tags:
 - InternVL2_5
 - InternVL2_5-1B
+- InternVL2_5-1B-MPO
 - Int8
 - VLM
 ---
@@ -36,16 +38,24 @@ https://huggingface.co/OpenGVLab/InternVL2_5-1B
   - [M4N-Dock(爱芯派Pro)](https://wiki.sipeed.com/hardware/zh/maixIV/m4ndock/m4ndock.html)
   - [M.2 Accelerator card](https://axcl-docs.readthedocs.io/zh-cn/latest/doc_guide_hardware.html)
 - AX630C
-  - *developing*
+  - [爱芯派2](https://axera-pi-2-docs-cn.readthedocs.io/zh-cn/latest/index.html)
+  - [Module-LLM](https://docs.m5stack.com/zh_CN/module/Module-LLM)
+  - [LLM630 Compute Kit](https://docs.m5stack.com/zh_CN/core/LLM630%20Compute%20Kit)
  
 |Chips|image encoder 448|ttft|w8a16|
 |--|--|--|--|
 |AX650| 350 ms | 420 ms |32 tokens/sec|
 
+|Chips|image encoder 364|ttft|w8a16|
+|--|--|--|--|
+|AX63C| 350 ms | 1100 ms |11 tokens/sec|
+
+
 ## How to use
 
 Download all files from this repository to the device
 
+**If you using AX650 Board**
 ```
 root@ax650:/mnt/qtang/llm-test/temp/InternVL2_5-1B# tree -L 1
 .
@@ -58,6 +68,18 @@ root@ax650:/mnt/qtang/llm-test/temp/InternVL2_5-1B# tree -L 1
 `-- ssd_car.jpg
 ```
 
+**If you using AX630C Board**
+```
+root@ax630c:/mnt/qtang/llm-test/internvl2_5-1b-mpo# tree -L 1
+.
+|-- config.json
+|-- internvl2_5_1b_364_ax630c
+|-- internvl2_5_tokenizer
+|-- internvl2_5_tokenizer_364.py
+|-- main
+`-- run_internvl2_5_364_ax630c.sh
+```
+
 #### Install transformer
 
 ```
@@ -66,6 +88,7 @@ pip install transformers==4.41.1
 
 #### Start the Tokenizer service
 
+**If you using AX650 Board**
 ```
 root@ax650:/mnt/qtang/llm-test/temp/InternVL2_5-1B# python3 internvl2_5_tokenizer_448.py --port 12345
 None None 151645 <|im_end|>
@@ -85,6 +108,25 @@ None None 151645 <|im_end|>
 http://localhost:12345
 ```
 
+**If you using AX630C Board**
+```
+root@ax650:/mnt/qtang/llm-test/temp/InternVL2_5-1B# python internvl2_5_tokenizer_364.py --port 12345
+Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
+None None 151645 <|im_end|>
+[151644, 8948, 198, 56568, 104625, 100633, 104455, 104800, 101101, 32022, 102022, 99602, 100013, 9370, 90286, 21287,
+42140, 53772, 35243, 26288, 104949, 3837, 105205, 109641, 67916, 30698, 11, 54851, 46944, 115404, 42192, 99441, 100623,
+48692, 100168, 110498, 1773, 151645, 151644, 872, 198, 151665, 151667, 151667, 151667, 151667, 151667, 151667, 151667,
+......
+151667, 151667, 151667, 151667, 151667, 151667, 151667, 151667, 151667, 151667, 151667, 151667, 151666, 198, 5501, 7512,
+279, 2168, 19620, 13, 151645, 151644, 77091, 198]
+223
+[151644, 8948, 198, 56568, 104625, 100633, 104455, 104800, 101101, 32022, 102022, 99602, 100013, 9370, 90286, 21287, 42140,
+53772, 35243, 26288, 104949, 3837, 105205, 109641, 67916, 30698, 11, 54851, 46944, 115404, 42192, 99441, 100623, 48692,
+100168, 110498, 1773, 151645, 151644, 872, 198, 14990, 1879, 151645, 151644, 77091, 198]
+47
+http://localhost:12345
+```
+
 #### Inference with AX650 Host, such as M4N-Dock(爱芯派Pro) or AX650N DEMO Board
 
 - input text
@@ -105,7 +147,7 @@ root@ax650:/mnt/qtang/llm-test/temp/InternVL2_5-1B# ./run_internvl2_5_448_ax650.
 bos_id: -1, eos_id: 151645
   3% | ██                                |   1 /  28 [0.01s<0.14s, 200.00 count/s] tokenizer init ok
 [I][                            Init][  26]: LLaMaEmbedSelector use mmap
-100% | ████████████████████████████████ |  28 /  28 [1.42s<1.42s, 19.66 count/s] init vpm axmodel ok,remain_cmm(2859 MB)B)
+100% | ████████████████████████████████ |  28 /  28 [1.42s<1.42s, 19.66 count/s] init vpm axmodel ok,remain_cmm(2859 MB)
 [I][                            Init][ 275]: max_token_len : 1023
 [I][                            Init][ 280]: kv_cache_size : 128, kv_cache_num: 1023
 [I][                            Init][ 288]: prefill_token_num : 320
@@ -134,4 +176,45 @@ root@ax650:/mnt/qtang/llm-test/temp/InternVL2_5-1B#
 
 [What is M.2 Accelerator card?](https://axcl-docs.readthedocs.io/zh-cn/latest/doc_guide_hardware.html), Show this DEMO based on Raspberry PI 5.
 
-*TODO*
+*TODO*
+
+#### Inference with AX630C Host, such as 爱芯派2, Module-LLM, LLM630 Compute Kit and AX630C DEMO Board
+
+- input text
+
+```
+Describe the picture
+```
+
+- input image
+
+![](./panda.jpg)
+
+Open another terminal and run `./run_internvl2_5_364_ax630c.sh`
+
+```
+/mnt/qtang/llm-test/internvl2_5-1b-mpo # ./run_internvl2_5_364_ax630c.sh
+[I][                            Init][ 106]: LLM init start
+bos_id: -1, eos_id: 151645
+  3% | ██                                |   1 /  28 [0.01s<0.14s, 200.00 count/s] tokenizer init ok
+[I][                            Init][  26]: LLaMaEmbedSelector use mmap
+100% | ████████████████████████████████ |  28 /  28 [9.48s<9.48s, 2.95 count/s] init vpm axmodel ok,remain_cmm(905 MB)
+[I][                            Init][ 254]: max_token_len : 1023
+[I][                            Init][ 259]: kv_cache_size : 128, kv_cache_num: 1023
+[I][                            Init][ 267]: prefill_token_num : 256
+[I][                            Init][ 269]: vpm_height : 364,vpm_width : 364
+[I][                            Init][ 278]: LLM init ok
+Type "q" to exit, Ctrl+c to stop current running
+
+prompt >> Please describe the image
+image >> panda.jpg
+[I][                          Encode][ 337]: image encode time : 1156.637939 ms, size : 151424
+[I][                             Run][ 548]: ttft: 1120.15 ms
+
+The image features a red panda in a natural setting, likely in a zoo or a forested area.
+The red panda has distinctive reddish-brown fur with white markings around its eyes and ears.
+It is leaning on a wooden structure, possibly a platform or a log, with a background of green foliage.
+The red panda appears to be looking directly at the camera with a calm expression.
+
+[N][                             Run][ 687]: hit eos,avg 10.94 token/s
+```