mirror of
https://www.modelscope.cn/moonshotai/Kimi-VL-A3B-Thinking-2506.git
synced 2026-04-02 19:13:01 +08:00
Upload folder using ModelScope SDK
This commit is contained in:
Cherrytest
165
.ipynb_checkpoints/README-checkpoint.md
Normal file
165
.ipynb_checkpoints/README-checkpoint.md
Normal file
@ -0,0 +1,165 @@
|
|||||||
|
---
|
||||||
|
base_model:
|
||||||
|
- moonshotai/Kimi-VL-A3B-Instruct
|
||||||
|
license: mit
|
||||||
|
pipeline_tag: image-text-to-text
|
||||||
|
library_name: transformers
|
||||||
|
---
|
||||||
|
|
||||||
|
> [!Note]
|
||||||
|
> This is an improved version of [Kimi-VL-A3B-Thinking](https://huggingface.co/moonshotai/Kimi-VL-A3B-Thinking). Please consider using this updated model instead of the previous version.
|
||||||
|
|
||||||
|
<div align="center">
|
||||||
|
<img width="30%" src="figures/logo.png">
|
||||||
|
</div>
|
||||||
|
|
||||||
|
<div align="center">
|
||||||
|
<a href="https://arxiv.org/abs/2504.07491">
|
||||||
|
<b>📄 Tech Report</b>
|
||||||
|
</a> |
|
||||||
|
<a href="https://github.com/MoonshotAI/Kimi-VL">
|
||||||
|
<b>📄 Github</b>
|
||||||
|
</a> |
|
||||||
|
<a href="https://huggingface.co/spaces/moonshotai/Kimi-VL-A3B-Thinking-2506/">💬 Chat Web</a>
|
||||||
|
</div>
|
||||||
|
|
||||||
|
## 1. Introduction
|
||||||
|
|
||||||
|
|
||||||
|
Two months after the initial release of our first open-source multimodal reasoning model, [Kimi-VL-A3B-Thinking](https://huggingface.co/moonshotai/Kimi-VL-A3B-Thinking), we update a regular improved version, [Kimi-VL-A3B-Thinking-2506](https://huggingface.co/moonshotai/Kimi-VL-A3B-Thinking-2506). Compared to the previous version, this new 2506 version provides several new or improved abilities:
|
||||||
|
|
||||||
|
- **It Thinks Smarter while Consuming Less Tokens**: The 2506 version reaches better accuracy on multimodal reasoning benchmarks: 56.9 on MathVision (+20.1), 80.1 on MathVista (+8.4), 46.2 on MMMU-Pro (+3.2), 64.0 on MMMU (+2.1), while in average requires 20\% reduced thinking length.
|
||||||
|
- **It Sees Clearer with Thinking**: Unlike the previous version that specializes on thinking tasks, the 2506 version can also achieve the same or even better ability on general visual perception and understanding, e.g. MMBench-EN-v1.1 (84.4), MMStar (70.4), RealWorldQA (70.0), MMVet (78.4), surpassing or matching abilties of our non-thinking model ([Kimi-VL-A3B-Instruct](https://huggingface.co/moonshotai/Kimi-VL-A3B-Instruct)).
|
||||||
|
- **It Extends to Video Scenarios**: The new 2506 version also improves on video reasoning and understanding benchmarks. It sets new state-of-the-art for open-source models on VideoMMMU (65.2), while also retains good ability on general video understanding (71.9 on Video-MME, matching [Kimi-VL-A3B-Instruct](https://huggingface.co/moonshotai/Kimi-VL-A3B-Instruct)).
|
||||||
|
- **It Extends to Higher Resolution**: The new 2506 version supports 3.2 million total pixels in a single image, 4X compared to the previous version. This leads to non-trivial improvements on high-resolution perception and OS-agent grounding benchmarks: 83.2 on V* Benchmark (without extra tools), 52.8 on ScreenSpot-Pro, 52.5 on OSWorld-G (full set with refusal).
|
||||||
|
|
||||||
|
In this blog, we provide simple examples to use this new model on image reasoning, video reasoning, and OS-agent scenarios.
|
||||||
|
|
||||||
|
## 2. Performance
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
## 3. Usage
|
||||||
|
|
||||||
|
### 3.1. Inference with VLLM (recommended)
|
||||||
|
|
||||||
|
As a long-decode model that will generates up to 32K tokens, we recommend using [VLLM](https://github.com/vllm-project/vllm/tree/main/vllm) for inference, which has already supported Kimi-VL series.
|
||||||
|
|
||||||
|
```shell
|
||||||
|
MAX_JOBS=4 pip install vllm==0.9.1 blobfile flash-attn --no-build-isolation
|
||||||
|
```
|
||||||
|
|
||||||
|
> [!Note]
|
||||||
|
> It is important to explicitly install flash-attn to avoid CUDA out-of-memory.
|
||||||
|
|
||||||
|
```python
|
||||||
|
from transformers import AutoProcessor
|
||||||
|
from vllm import LLM, SamplingParams
|
||||||
|
|
||||||
|
model_path = "moonshotai/Kimi-VL-A3B-Thinking-2506"
|
||||||
|
llm = LLM(
|
||||||
|
model_path,
|
||||||
|
trust_remote_code=True,
|
||||||
|
max_num_seqs=8,
|
||||||
|
max_model_len=131072,
|
||||||
|
limit_mm_per_prompt={"image": 256}
|
||||||
|
)
|
||||||
|
|
||||||
|
processor = AutoProcessor.from_pretrained(model_path, trust_remote_code=True)
|
||||||
|
|
||||||
|
sampling_params = SamplingParams(max_tokens=32768, temperature=0.8)
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
import requests
|
||||||
|
from PIL import Image
|
||||||
|
|
||||||
|
def extract_thinking_and_summary(text: str, bot: str = "◁think▷", eot: str = "◁/think▷") -> str:
|
||||||
|
if bot in text and eot not in text:
|
||||||
|
return ""
|
||||||
|
if eot in text:
|
||||||
|
return text[text.index(bot) + len(bot):text.index(eot)].strip(), text[text.index(eot) + len(eot) :].strip()
|
||||||
|
return "", text
|
||||||
|
|
||||||
|
OUTPUT_FORMAT = "--------Thinking--------\n{thinking}\n\n--------Summary--------\n{summary}"
|
||||||
|
|
||||||
|
url = "https://huggingface.co/spaces/moonshotai/Kimi-VL-A3B-Thinking/resolve/main/images/demo6.jpeg"
|
||||||
|
image = Image.open(requests.get(url,stream=True).raw)
|
||||||
|
|
||||||
|
messages = [
|
||||||
|
{"role": "user", "content": [{"type": "image", "image": ""}, {"type": "text", "text": "What kind of cat is this? Answer with one word."}]}
|
||||||
|
]
|
||||||
|
text = processor.apply_chat_template(messages, add_generation_prompt=True, return_tensors="pt")
|
||||||
|
|
||||||
|
outputs = llm.generate([{"prompt": text, "multi_modal_data": {"image": image}}], sampling_params=sampling_params)
|
||||||
|
generated_text = outputs[0].outputs[0].text
|
||||||
|
|
||||||
|
thinking, summary = extract_thinking_and_summary(generated_text)
|
||||||
|
print(OUTPUT_FORMAT.format(thinking=thinking, summary=summary))
|
||||||
|
```
|
||||||
|
|
||||||
|
|
||||||
|
### 3.2. Inference with 🤗 Hugging Face Transformers
|
||||||
|
|
||||||
|
We introduce how to use our model at inference stage using transformers library. It is recommended to use python=3.10, torch>=2.1.0, and transformers=4.48.2 as the development environment.
|
||||||
|
|
||||||
|
```python
|
||||||
|
from PIL import Image
|
||||||
|
from transformers import AutoModelForCausalLM, AutoProcessor
|
||||||
|
|
||||||
|
def extract_thinking_and_summary(text: str, bot: str = "◁think▷", eot: str = "◁/think▷") -> str:
|
||||||
|
if bot in text and eot not in text:
|
||||||
|
return ""
|
||||||
|
if eot in text:
|
||||||
|
return text[text.index(bot) + len(bot):text.index(eot)].strip(), text[text.index(eot) + len(eot) :].strip()
|
||||||
|
return "", text
|
||||||
|
|
||||||
|
OUTPUT_FORMAT = "--------Thinking--------\n{thinking}\n\n--------Summary--------\n{summary}"
|
||||||
|
|
||||||
|
url = "https://huggingface.co/spaces/moonshotai/Kimi-VL-A3B-Thinking/resolve/main/images/demo6.jpeg"
|
||||||
|
|
||||||
|
model_path = "moonshotai/Kimi-VL-A3B-Thinking-2506"
|
||||||
|
model = AutoModelForCausalLM.from_pretrained(
|
||||||
|
model_path,
|
||||||
|
torch_dtype="auto",
|
||||||
|
device_map="auto",
|
||||||
|
trust_remote_code=True,
|
||||||
|
)
|
||||||
|
processor = AutoProcessor.from_pretrained(model_path, trust_remote_code=True)
|
||||||
|
|
||||||
|
image_paths = ["url"]
|
||||||
|
images = [Image.open(path) for path in image_paths]
|
||||||
|
messages = [
|
||||||
|
{
|
||||||
|
"role": "user",
|
||||||
|
"content": [
|
||||||
|
{"type": "image", "image": image_path} for image_path in image_paths
|
||||||
|
] + [{"type": "text", "text": ""What kind of cat is this? Answer with one word."}],
|
||||||
|
},
|
||||||
|
]
|
||||||
|
text = processor.apply_chat_template(messages, add_generation_prompt=True, return_tensors="pt")
|
||||||
|
inputs = processor(images=images, text=text, return_tensors="pt", padding=True, truncation=True).to(model.device)
|
||||||
|
generated_ids = model.generate(**inputs, max_new_tokens=32768, temperature=0.8)
|
||||||
|
generated_ids_trimmed = [
|
||||||
|
out_ids[len(in_ids) :] for in_ids, out_ids in zip(inputs.input_ids, generated_ids)
|
||||||
|
]
|
||||||
|
response = processor.batch_decode(
|
||||||
|
generated_ids_trimmed, skip_special_tokens=True, clean_up_tokenization_spaces=False
|
||||||
|
)[0]
|
||||||
|
print(response)
|
||||||
|
```
|
||||||
|
|
||||||
|
|
||||||
|
## 4. Citation
|
||||||
|
|
||||||
|
```
|
||||||
|
@misc{kimiteam2025kimivltechnicalreport,
|
||||||
|
title={{Kimi-VL} Technical Report},
|
||||||
|
author={Kimi Team and Angang Du and Bohong Yin and Bowei Xing and Bowen Qu and Bowen Wang and Cheng Chen and Chenlin Zhang and Chenzhuang Du and Chu Wei and Congcong Wang and Dehao Zhang and Dikang Du and Dongliang Wang and Enming Yuan and Enzhe Lu and Fang Li and Flood Sung and Guangda Wei and Guokun Lai and Han Zhu and Hao Ding and Hao Hu and Hao Yang and Hao Zhang and Haoning Wu and Haotian Yao and Haoyu Lu and Heng Wang and Hongcheng Gao and Huabin Zheng and Jiaming Li and Jianlin Su and Jianzhou Wang and Jiaqi Deng and Jiezhong Qiu and Jin Xie and Jinhong Wang and Jingyuan Liu and Junjie Yan and Kun Ouyang and Liang Chen and Lin Sui and Longhui Yu and Mengfan Dong and Mengnan Dong and Nuo Xu and Pengyu Cheng and Qizheng Gu and Runjie Zhou and Shaowei Liu and Sihan Cao and Tao Yu and Tianhui Song and Tongtong Bai and Wei Song and Weiran He and Weixiao Huang and Weixin Xu and Xiaokun Yuan and Xingcheng Yao and Xingzhe Wu and Xinxing Zu and Xinyu Zhou and Xinyuan Wang and Y. Charles and Yan Zhong and Yang Li and Yangyang Hu and Yanru Chen and Yejie Wang and Yibo Liu and Yibo Miao and Yidao Qin and Yimin Chen and Yiping Bao and Yiqin Wang and Yongsheng Kang and Yuanxin Liu and Yulun Du and Yuxin Wu and Yuzhi Wang and Yuzi Yan and Zaida Zhou and Zhaowei Li and Zhejun Jiang and Zheng Zhang and Zhilin Yang and Zhiqi Huang and Zihao Huang and Zijia Zhao and Ziwei Chen},
|
||||||
|
year={2025},
|
||||||
|
eprint={2504.07491},
|
||||||
|
archivePrefix={arXiv},
|
||||||
|
primaryClass={cs.CV},
|
||||||
|
url={https://arxiv.org/abs/2504.07491},
|
||||||
|
}
|
||||||
|
```
|
||||||
253
README.md
253
README.md
@ -1,47 +1,220 @@
|
|||||||
---
|
---
|
||||||
license: Apache License 2.0
|
base_model:
|
||||||
|
- moonshotai/Kimi-VL-A3B-Instruct
|
||||||
#model-type:
|
license: mit
|
||||||
##如 gpt、phi、llama、chatglm、baichuan 等
|
pipeline_tag: image-text-to-text
|
||||||
#- gpt
|
library_name: transformers
|
||||||
|
|
||||||
#domain:
|
|
||||||
##如 nlp、cv、audio、multi-modal
|
|
||||||
#- nlp
|
|
||||||
|
|
||||||
#language:
|
|
||||||
##语言代码列表 https://help.aliyun.com/document_detail/215387.html?spm=a2c4g.11186623.0.0.9f8d7467kni6Aa
|
|
||||||
#- cn
|
|
||||||
|
|
||||||
#metrics:
|
|
||||||
##如 CIDEr、Blue、ROUGE 等
|
|
||||||
#- CIDEr
|
|
||||||
|
|
||||||
#tags:
|
|
||||||
##各种自定义，包括 pretrained、fine-tuned、instruction-tuned、RL-tuned 等训练方法和其他
|
|
||||||
#- pretrained
|
|
||||||
|
|
||||||
#tools:
|
|
||||||
##如 vllm、fastchat、llamacpp、AdaSeq 等
|
|
||||||
#- vllm
|
|
||||||
---
|
---
|
||||||
### 当前模型的贡献者未提供更加详细的模型介绍。模型文件和权重，可浏览“模型文件”页面获取。
|
|
||||||
#### 您可以通过如下git clone命令，或者ModelScope SDK来下载模型
|
|
||||||
|
|
||||||
SDK下载
|
> [!Note]
|
||||||
```bash
|
> This is an improved version of [Kimi-VL-A3B-Thinking](https://huggingface.co/moonshotai/Kimi-VL-A3B-Thinking). Please consider using this updated model instead of the previous version.
|
||||||
#安装ModelScope
|
|
||||||
pip install modelscope
|
<div align="center">
|
||||||
|
<img width="80%" src="figures/logo.png">
|
||||||
|
</div>
|
||||||
|
|
||||||
|
<div align="center">
|
||||||
|
<a href="https://arxiv.org/abs/2504.07491">
|
||||||
|
<b>📄 Tech Report</b>
|
||||||
|
</a> |
|
||||||
|
<a href="https://github.com/MoonshotAI/Kimi-VL">
|
||||||
|
<b>📄 Github</b>
|
||||||
|
</a> |
|
||||||
|
<a href="https://huggingface.co/spaces/moonshotai/Kimi-VL-A3B-Thinking-2506/">💬 Chat Web</a>
|
||||||
|
</div>
|
||||||
|
|
||||||
|
## 1. Introduction
|
||||||
|
|
||||||
|
This is an updated version of [Kimi-VL-A3B-Thinking](https://huggingface.co/moonshotai/Kimi-VL-A3B-Thinking), with following improved abilities:
|
||||||
|
|
||||||
|
- **It Thinks Smarter while Consuming Less Tokens**: The 2506 version reaches better accuracy on multimodal reasoning benchmarks: 56.9 on MathVision (+20.1), 80.1 on MathVista (+8.4), 46.3 on MMMU-Pro (+3.3), 64.0 on MMMU (+2.1), while in average requires 20\% reduced thinking length.
|
||||||
|
- **It Sees Clearer with Thinking**: Unlike the previous version that specializes on thinking tasks, the 2506 version can also achieve the same or even better ability on general visual perception and understanding, e.g. MMBench-EN-v1.1 (84.4), MMStar (70.4), RealWorldQA (70.0), MMVet (78.4), surpassing or matching abilties of our non-thinking model ([Kimi-VL-A3B-Instruct](https://huggingface.co/moonshotai/Kimi-VL-A3B-Instruct)).
|
||||||
|
- **It Extends to Video Scenarios**: The new 2506 version also improves on video reasoning and understanding benchmarks. It sets new state-of-the-art for open-source models on VideoMMMU (65.2), while also retains good ability on general video understanding (71.9 on Video-MME, matching [Kimi-VL-A3B-Instruct](https://huggingface.co/moonshotai/Kimi-VL-A3B-Instruct)).
|
||||||
|
- **It Extends to Higher Resolution**: The new 2506 version supports 3.2 million total pixels in a single image, 4X compared to the previous version. This leads to non-trivial improvements on high-resolution perception and OS-agent grounding benchmarks: 83.2 on V* Benchmark (without extra tools), 52.8 on ScreenSpot-Pro, 52.5 on OSWorld-G (full set with refusal).
|
||||||
|
|
||||||
|
|
||||||
|
## 2. Performance
|
||||||
|
|
||||||
|
Comparison with efficient models and two previous versions of Kimi-VL:
|
||||||
|
|
||||||
|
<div align="center">
|
||||||
|
|
||||||
|
| Benchmark (Metric) | GPT-4o | Qwen2.5-VL-7B | Gemma3-12B-IT | Kimi-VL-A3B-Instruct | Kimi-VL-A3B-Thinking | Kimi-VL-A3B-Thinking-2506 |
|
||||||
|
|----------------------------|--------|---------------|---------------|----------------------|----------------------|--------------------------|
|
||||||
|
| **General Multimodal** | | | | | | |
|
||||||
|
| MMBench-EN-v1.1 (Acc) | 83.1 | 83.2 | 74.6 | 82.9 | 76.0 | **84.4** |
|
||||||
|
| RealWorldQA (Acc) | 75.4 | 68.5 | 59.1 | 68.1 | 64.0 | **70.0** |
|
||||||
|
| OCRBench (Acc) | 815 | 864 | 702 | 864 | 864 | **869** |
|
||||||
|
| MMStar (Acc) | 64.7 | 63.0 | 56.1 | 61.7 | 64.2 | **70.4** |
|
||||||
|
| MMVet (Acc) | 69.1 | 67.1 | 64.9 | 66.7 | 69.5 | **78.1** |
|
||||||
|
| **Reasoning** | | | | | | |
|
||||||
|
| MMMU (val, Pass@1) | 69.1 | 58.6 | 59.6 | 57.0 | 61.7 | **64.0** |
|
||||||
|
| MMMU-Pro (Pass@1) | 51.7 | 38.1 | 32.1 | 36.0 | 43.2 | **46.3** |
|
||||||
|
| **Math** | | | | | | |
|
||||||
|
| MATH-Vision (Pass@1) | 30.4 | 25.0 | 32.1 | 21.7 | 36.8 | **56.9** |
|
||||||
|
| MathVista_MINI (Pass@1) | 63.8 | 68.0 | 56.1 | 68.6 | 71.7 | **80.1** |
|
||||||
|
| **Video** | | | | | | |
|
||||||
|
| VideoMMMU (Pass@1) | 61.2 | 47.4 | 57.0 | 52.1 | 55.5 | **65.2** |
|
||||||
|
| MMVU (Pass@1) | 67.4 | 50.1 | 57.0 | 52.7 | 53.0 | **57.5** |
|
||||||
|
| Video-MME (w/ sub.) | 77.2 | 71.6 | 62.1 | **72.7** | 66.0 | 71.9 |
|
||||||
|
| **Agent Grounding** | | | | | | |
|
||||||
|
| ScreenSpot-Pro (Acc) | 0.8 | 29.0 | — | 35.4 | — | **52.8** |
|
||||||
|
| ScreenSpot-V2 (Acc) | 18.1 | 84.2 | — | **92.8** | — | 91.4 |
|
||||||
|
| OSWorld-G (Acc) | - | 31.5 | — | 41.6 | — | **52.5** |
|
||||||
|
| **Long Document** | | | | | | |
|
||||||
|
| MMLongBench-DOC (Acc) | 42.8 | 29.6 | 21.3 | 35.1 | 32.5 | **42.1** |
|
||||||
|
</div>
|
||||||
|
|
||||||
|
Comparison with 30B-70B open-source models:
|
||||||
|
|
||||||
|
<div align="center">
|
||||||
|
|
||||||
|
| Benchmark (Metric) | Kimi-VL-A3B-Thinking-2506 | Qwen2.5-VL-32B | Qwen2.5-VL-72B | Gemma3-27B-IT |
|
||||||
|
|----------------------------|---------------------------|---------------|---------------|---------------|
|
||||||
|
| **General Multimodal** | | | | |
|
||||||
|
| MMBench-EN-v1.1 (Acc) | 84.4 | - | 88.3 | 78.9 |
|
||||||
|
| RealWorldQA (Acc) | 70.0 | - | 75.7 | 62.5 |
|
||||||
|
| OCRBench (Acc) | 869 | - | 885 | 753 |
|
||||||
|
| MMStar (Acc) | 70.4 | 69.5 | 70.8 | 63.1 |
|
||||||
|
| MMVet (Acc) | 78.1 | - | 74.0 | 71.0 |
|
||||||
|
| **Reasoning** | | | ||
|
||||||
|
| MMMU (val, Pass@1) | 64.0 | 70.0 | 70.2 | 64.9 |
|
||||||
|
| MMMU-Pro (Pass@1) | 46.3 | 49.5 | 51.1 | - |
|
||||||
|
| MATH-Vision (Pass@1) | 56.9 | 38.4 | 38.1 | 35.4 |
|
||||||
|
| MathVista\_MINI (Pass@1) | 80.1 | 74.7 | 74.8 | 59.8 |
|
||||||
|
| **Video** | | | | |
|
||||||
|
| VideoMMMU (Pass@1) | 65.2 | - | 60.2 | 61.8 |
|
||||||
|
| MMVU (Pass@1) | 57.5 | - | 62.9 | 61.3 |
|
||||||
|
| Video-MME (w/ sub.) | 71.9 | 70.5/77.9 | 73.3/79.1 | - |
|
||||||
|
| **Agent Grounding** | | | | |
|
||||||
|
| ScreenSpot-Pro (Acc) | 52.8 | 39.4 | 43.6 | - |
|
||||||
|
| ScreenSpot-V2 (Acc) | 91.4 | - | - | - |
|
||||||
|
| OSWorld-G (Acc) | 52.5 | 46.5 | - | - |
|
||||||
|
| **Long Document** | | | | |
|
||||||
|
| MMLongBench-DOC (Acc) | 42.1 | - | 38.8 | - |
|
||||||
|
</div>
|
||||||
|
|
||||||
|
|
||||||
|
## 3. Usage
|
||||||
|
|
||||||
|
### 3.1. Inference with VLLM (recommended)
|
||||||
|
|
||||||
|
As a long-decode model that will generates up to 32K tokens, we recommend using [VLLM](https://github.com/vllm-project/vllm/tree/main/vllm) for inference, which has already supported Kimi-VL series.
|
||||||
|
|
||||||
|
```shell
|
||||||
|
MAX_JOBS=4 pip install vllm==0.9.1 blobfile flash-attn --no-build-isolation
|
||||||
```
|
```
|
||||||
|
|
||||||
|
> [!Note]
|
||||||
|
> It is important to explicitly install flash-attn to avoid CUDA out-of-memory.
|
||||||
|
|
||||||
```python
|
```python
|
||||||
#SDK模型下载
|
from transformers import AutoProcessor
|
||||||
from modelscope import snapshot_download
|
from vllm import LLM, SamplingParams
|
||||||
model_dir = snapshot_download('moonshotai/Kimi-VL-A3B-Thinking-2506')
|
|
||||||
```
|
model_path = "moonshotai/Kimi-VL-A3B-Thinking-2506"
|
||||||
Git下载
|
llm = LLM(
|
||||||
```
|
model_path,
|
||||||
#Git模型下载
|
trust_remote_code=True,
|
||||||
git clone https://www.modelscope.cn/moonshotai/Kimi-VL-A3B-Thinking-2506.git
|
max_num_seqs=8,
|
||||||
|
max_model_len=131072,
|
||||||
|
limit_mm_per_prompt={"image": 256}
|
||||||
|
)
|
||||||
|
|
||||||
|
processor = AutoProcessor.from_pretrained(model_path, trust_remote_code=True)
|
||||||
|
|
||||||
|
sampling_params = SamplingParams(max_tokens=32768, temperature=0.8)
|
||||||
|
|
||||||
|
|
||||||
|
import requests
|
||||||
|
from PIL import Image
|
||||||
|
|
||||||
|
def extract_thinking_and_summary(text: str, bot: str = "◁think▷", eot: str = "◁/think▷") -> str:
|
||||||
|
if bot in text and eot not in text:
|
||||||
|
return ""
|
||||||
|
if eot in text:
|
||||||
|
return text[text.index(bot) + len(bot):text.index(eot)].strip(), text[text.index(eot) + len(eot) :].strip()
|
||||||
|
return "", text
|
||||||
|
|
||||||
|
OUTPUT_FORMAT = "--------Thinking--------\n{thinking}\n\n--------Summary--------\n{summary}"
|
||||||
|
|
||||||
|
url = "https://huggingface.co/spaces/moonshotai/Kimi-VL-A3B-Thinking/resolve/main/images/demo6.jpeg"
|
||||||
|
image = Image.open(requests.get(url,stream=True).raw)
|
||||||
|
|
||||||
|
messages = [
|
||||||
|
{"role": "user", "content": [{"type": "image", "image": ""}, {"type": "text", "text": "What kind of cat is this? Answer with one word."}]}
|
||||||
|
]
|
||||||
|
text = processor.apply_chat_template(messages, add_generation_prompt=True, return_tensors="pt")
|
||||||
|
|
||||||
|
outputs = llm.generate([{"prompt": text, "multi_modal_data": {"image": image}}], sampling_params=sampling_params)
|
||||||
|
generated_text = outputs[0].outputs[0].text
|
||||||
|
|
||||||
|
thinking, summary = extract_thinking_and_summary(generated_text)
|
||||||
|
print(OUTPUT_FORMAT.format(thinking=thinking, summary=summary))
|
||||||
```
|
```
|
||||||
|
|
||||||
<p style="color: lightgrey;">如果您是本模型的贡献者，我们邀请您根据<a href="https://modelscope.cn/docs/ModelScope%E6%A8%A1%E5%9E%8B%E6%8E%A5%E5%85%A5%E6%B5%81%E7%A8%8B%E6%A6%82%E8%A7%88" style="color: lightgrey; text-decoration: underline;">模型贡献文档</a>，及时完善模型卡片内容。</p>
|
|
||||||
|
### 3.2. Inference with 🤗 Hugging Face Transformers
|
||||||
|
|
||||||
|
We introduce how to use our model at inference stage using transformers library. It is recommended to use python=3.10, torch>=2.1.0, and transformers=4.48.2 as the development environment.
|
||||||
|
|
||||||
|
```python
|
||||||
|
from PIL import Image
|
||||||
|
from transformers import AutoModelForCausalLM, AutoProcessor
|
||||||
|
|
||||||
|
def extract_thinking_and_summary(text: str, bot: str = "◁think▷", eot: str = "◁/think▷") -> str:
|
||||||
|
if bot in text and eot not in text:
|
||||||
|
return ""
|
||||||
|
if eot in text:
|
||||||
|
return text[text.index(bot) + len(bot):text.index(eot)].strip(), text[text.index(eot) + len(eot) :].strip()
|
||||||
|
return "", text
|
||||||
|
|
||||||
|
OUTPUT_FORMAT = "--------Thinking--------\n{thinking}\n\n--------Summary--------\n{summary}"
|
||||||
|
|
||||||
|
url = "https://huggingface.co/spaces/moonshotai/Kimi-VL-A3B-Thinking/resolve/main/images/demo6.jpeg"
|
||||||
|
|
||||||
|
model_path = "moonshotai/Kimi-VL-A3B-Thinking-2506"
|
||||||
|
model = AutoModelForCausalLM.from_pretrained(
|
||||||
|
model_path,
|
||||||
|
torch_dtype="auto",
|
||||||
|
device_map="auto",
|
||||||
|
trust_remote_code=True,
|
||||||
|
)
|
||||||
|
processor = AutoProcessor.from_pretrained(model_path, trust_remote_code=True)
|
||||||
|
|
||||||
|
image_paths = ["url"]
|
||||||
|
images = [Image.open(path) for path in image_paths]
|
||||||
|
messages = [
|
||||||
|
{
|
||||||
|
"role": "user",
|
||||||
|
"content": [
|
||||||
|
{"type": "image", "image": image_path} for image_path in image_paths
|
||||||
|
] + [{"type": "text", "text": ""What kind of cat is this? Answer with one word."}],
|
||||||
|
},
|
||||||
|
]
|
||||||
|
text = processor.apply_chat_template(messages, add_generation_prompt=True, return_tensors="pt")
|
||||||
|
inputs = processor(images=images, text=text, return_tensors="pt", padding=True, truncation=True).to(model.device)
|
||||||
|
generated_ids = model.generate(**inputs, max_new_tokens=32768, temperature=0.8)
|
||||||
|
generated_ids_trimmed = [
|
||||||
|
out_ids[len(in_ids) :] for in_ids, out_ids in zip(inputs.input_ids, generated_ids)
|
||||||
|
]
|
||||||
|
response = processor.batch_decode(
|
||||||
|
generated_ids_trimmed, skip_special_tokens=True, clean_up_tokenization_spaces=False
|
||||||
|
)[0]
|
||||||
|
print(response)
|
||||||
|
```
|
||||||
|
|
||||||
|
|
||||||
|
## 4. Citation
|
||||||
|
|
||||||
|
```
|
||||||
|
@misc{kimiteam2025kimivltechnicalreport,
|
||||||
|
title={{Kimi-VL} Technical Report},
|
||||||
|
author={Kimi Team and Angang Du and Bohong Yin and Bowei Xing and Bowen Qu and Bowen Wang and Cheng Chen and Chenlin Zhang and Chenzhuang Du and Chu Wei and Congcong Wang and Dehao Zhang and Dikang Du and Dongliang Wang and Enming Yuan and Enzhe Lu and Fang Li and Flood Sung and Guangda Wei and Guokun Lai and Han Zhu and Hao Ding and Hao Hu and Hao Yang and Hao Zhang and Haoning Wu and Haotian Yao and Haoyu Lu and Heng Wang and Hongcheng Gao and Huabin Zheng and Jiaming Li and Jianlin Su and Jianzhou Wang and Jiaqi Deng and Jiezhong Qiu and Jin Xie and Jinhong Wang and Jingyuan Liu and Junjie Yan and Kun Ouyang and Liang Chen and Lin Sui and Longhui Yu and Mengfan Dong and Mengnan Dong and Nuo Xu and Pengyu Cheng and Qizheng Gu and Runjie Zhou and Shaowei Liu and Sihan Cao and Tao Yu and Tianhui Song and Tongtong Bai and Wei Song and Weiran He and Weixiao Huang and Weixin Xu and Xiaokun Yuan and Xingcheng Yao and Xingzhe Wu and Xinxing Zu and Xinyu Zhou and Xinyuan Wang and Y. Charles and Yan Zhong and Yang Li and Yangyang Hu and Yanru Chen and Yejie Wang and Yibo Liu and Yibo Miao and Yidao Qin and Yimin Chen and Yiping Bao and Yiqin Wang and Yongsheng Kang and Yuanxin Liu and Yulun Du and Yuxin Wu and Yuzhi Wang and Yuzi Yan and Zaida Zhou and Zhaowei Li and Zhejun Jiang and Zheng Zhang and Zhilin Yang and Zhiqi Huang and Zihao Huang and Zijia Zhao and Ziwei Chen},
|
||||||
|
year={2025},
|
||||||
|
eprint={2504.07491},
|
||||||
|
archivePrefix={arXiv},
|
||||||
|
primaryClass={cs.CV},
|
||||||
|
url={https://arxiv.org/abs/2504.07491},
|
||||||
|
}
|
||||||
|
```
|
||||||
31
chat_template.jinja
Normal file
31
chat_template.jinja
Normal file
@ -0,0 +1,31 @@
|
|||||||
|
{%- for message in messages -%}
|
||||||
|
{%- if loop.first and messages[0]['role'] != 'system' -%}
|
||||||
|
{{'<|im_system|>system<|im_middle|>You are a helpful assistant<|im_end|>'}}
|
||||||
|
{%- endif -%}
|
||||||
|
{%- if message['role'] == 'system' -%}
|
||||||
|
{{'<|im_system|>'}}
|
||||||
|
{%- endif -%}
|
||||||
|
{%- if message['role'] == 'user' -%}
|
||||||
|
{{'<|im_user|>'}}
|
||||||
|
{%- endif -%}
|
||||||
|
{%- if message['role'] == 'assistant' -%}
|
||||||
|
{{'<|im_assistant|>'}}
|
||||||
|
{%- endif -%}
|
||||||
|
{{- message['role'] -}}
|
||||||
|
{{'<|im_middle|>'}}
|
||||||
|
{%- if message['content'] is string -%}
|
||||||
|
{{- message['content'] + '<|im_end|>' -}}
|
||||||
|
{%- else -%}
|
||||||
|
{%- for content in message['content'] -%}
|
||||||
|
{%- if content['type'] == 'image' or 'image' in content or 'image_url' in content -%}
|
||||||
|
{{'<|media_start|>image<|media_content|><|media_pad|><|media_end|>'}}
|
||||||
|
{%- else -%}
|
||||||
|
{{content['text']}}
|
||||||
|
{%- endif -%}
|
||||||
|
{%- endfor -%}
|
||||||
|
{{'<|im_end|>'}}
|
||||||
|
{%- endif -%}
|
||||||
|
{%- endfor -%}
|
||||||
|
{%- if add_generation_prompt -%}
|
||||||
|
{{'<|im_assistant|>assistant<|im_middle|>'}}
|
||||||
|
{%- endif -%}
|
||||||
75
config.json
Normal file
75
config.json
Normal file
@ -0,0 +1,75 @@
|
|||||||
|
{
|
||||||
|
"architectures": [
|
||||||
|
"KimiVLForConditionalGeneration"
|
||||||
|
],
|
||||||
|
"auto_map": {
|
||||||
|
"AutoConfig": "configuration_kimi_vl.KimiVLConfig",
|
||||||
|
"AutoModel": "modeling_kimi_vl.KimiVLForConditionalGeneration",
|
||||||
|
"AutoModelForCausalLM": "modeling_kimi_vl.KimiVLForConditionalGeneration"
|
||||||
|
},
|
||||||
|
"vision_config": {
|
||||||
|
"model_type": "moonvit",
|
||||||
|
"patch_size": 14,
|
||||||
|
"num_attention_heads": 16,
|
||||||
|
"num_hidden_layers": 27,
|
||||||
|
"hidden_size": 1152,
|
||||||
|
"intermediate_size": 4304,
|
||||||
|
"init_pos_emb_height": 64,
|
||||||
|
"init_pos_emb_width": 64,
|
||||||
|
"merge_kernel_size": [
|
||||||
|
2,
|
||||||
|
2
|
||||||
|
],
|
||||||
|
"torch_dtype": "bfloat16"
|
||||||
|
},
|
||||||
|
"text_config": {
|
||||||
|
"vocab_size": 163840,
|
||||||
|
"max_position_embeddings": 131072,
|
||||||
|
"hidden_size": 2048,
|
||||||
|
"intermediate_size": 11264,
|
||||||
|
"moe_intermediate_size": 1408,
|
||||||
|
"num_hidden_layers": 27,
|
||||||
|
"num_attention_heads": 16,
|
||||||
|
"n_shared_experts": 2,
|
||||||
|
"n_routed_experts": 64,
|
||||||
|
"ep_size": 1,
|
||||||
|
"routed_scaling_factor": 2.446,
|
||||||
|
"kv_lora_rank": 512,
|
||||||
|
"q_lora_rank": null,
|
||||||
|
"qk_rope_head_dim": 64,
|
||||||
|
"v_head_dim": 128,
|
||||||
|
"qk_nope_head_dim": 128,
|
||||||
|
"topk_method": "noaux_tc",
|
||||||
|
"n_group": 1,
|
||||||
|
"topk_group": 1,
|
||||||
|
"num_experts_per_tok": 6,
|
||||||
|
"moe_layer_freq": 1,
|
||||||
|
"first_k_dense_replace": 1,
|
||||||
|
"norm_topk_prob": true,
|
||||||
|
"scoring_func": "sigmoid",
|
||||||
|
"aux_loss_alpha": 0.001,
|
||||||
|
"seq_aux": true,
|
||||||
|
"num_key_value_heads": 16,
|
||||||
|
"hidden_act": "silu",
|
||||||
|
"initializer_range": 0.02,
|
||||||
|
"rms_norm_eps": 1e-05,
|
||||||
|
"pretraining_tp": 1,
|
||||||
|
"use_cache": true,
|
||||||
|
"rope_theta": 800000.0,
|
||||||
|
"rope_scaling": null,
|
||||||
|
"attention_bias": false,
|
||||||
|
"attention_dropout": 0.0,
|
||||||
|
"bos_token_id": 163584,
|
||||||
|
"pad_token_id": 163839,
|
||||||
|
"eos_token_id": 163585,
|
||||||
|
"torch_dtype": "bfloat16",
|
||||||
|
"tie_word_embeddings": false
|
||||||
|
},
|
||||||
|
"ignore_index": -100,
|
||||||
|
"media_placeholder_token_id": 163605,
|
||||||
|
"torch_dtype": "bfloat16",
|
||||||
|
"transformers_version": "4.50.3",
|
||||||
|
"tie_word_embeddings": false,
|
||||||
|
"vocab_size": 163840,
|
||||||
|
"model_type": "kimi_vl"
|
||||||
|
}
|
||||||
1
configuration.json
Normal file
1
configuration.json
Normal file
@ -0,0 +1 @@
|
|||||||
|
{"framework": "pytorch", "task": "image-text-to-text", "allow_remote": true}
|
||||||
284
configuration_kimi_vl.py
Normal file
284
configuration_kimi_vl.py
Normal file
@ -0,0 +1,284 @@
|
|||||||
|
from transformers.configuration_utils import PretrainedConfig
|
||||||
|
from transformers.utils import logging
|
||||||
|
from typing import Optional, Union
|
||||||
|
|
||||||
|
logger = logging.get_logger(__name__)
|
||||||
|
|
||||||
|
DEEPSEEK_PRETRAINED_CONFIG_ARCHIVE_MAP = {}
|
||||||
|
|
||||||
|
|
||||||
|
class DeepseekV3Config(PretrainedConfig):
|
||||||
|
r"""
|
||||||
|
This is the configuration class to store the configuration of a [`DeepseekV3Model`]. It is used to instantiate an DeepSeek
|
||||||
|
model according to the specified arguments, defining the model architecture. Instantiating a configuration with the
|
||||||
|
defaults will yield a similar configuration to that of the DeepSeek-V3.
|
||||||
|
|
||||||
|
Configuration objects inherit from [`PretrainedConfig`] and can be used to control the model outputs. Read the
|
||||||
|
documentation from [`PretrainedConfig`] for more information.
|
||||||
|
|
||||||
|
Copy from https://huggingface.co/deepseek-ai/DeepSeek-V3/blob/main/configuration_deepseek.py
|
||||||
|
|
||||||
|
Args:
|
||||||
|
vocab_size (`int`, *optional*, defaults to 129280):
|
||||||
|
Vocabulary size of the Deep model. Defines the number of different tokens that can be represented by the
|
||||||
|
`inputs_ids` passed when calling [`DeepseekV3Model`]
|
||||||
|
hidden_size (`int`, *optional*, defaults to 4096):
|
||||||
|
Dimension of the hidden representations.
|
||||||
|
intermediate_size (`int`, *optional*, defaults to 11008):
|
||||||
|
Dimension of the MLP representations.
|
||||||
|
moe_intermediate_size (`int`, *optional*, defaults to 1407):
|
||||||
|
Dimension of the MoE representations.
|
||||||
|
num_hidden_layers (`int`, *optional*, defaults to 32):
|
||||||
|
Number of hidden layers in the Transformer decoder.
|
||||||
|
num_nextn_predict_layers (`int`, *optional*, defaults to 1):
|
||||||
|
Number of nextn predict layers in the DeepSeekV3 Model.
|
||||||
|
num_attention_heads (`int`, *optional*, defaults to 32):
|
||||||
|
Number of attention heads for each attention layer in the Transformer decoder.
|
||||||
|
n_shared_experts (`int`, *optional*, defaults to None):
|
||||||
|
Number of shared experts, None means dense model.
|
||||||
|
n_routed_experts (`int`, *optional*, defaults to None):
|
||||||
|
Number of routed experts, None means dense model.
|
||||||
|
routed_scaling_factor (`float`, *optional*, defaults to 1.0):
|
||||||
|
Scaling factor or routed experts.
|
||||||
|
topk_method (`str`, *optional*, defaults to `gready`):
|
||||||
|
Topk method used in routed gate.
|
||||||
|
n_group (`int`, *optional*, defaults to None):
|
||||||
|
Number of groups for routed experts.
|
||||||
|
topk_group (`int`, *optional*, defaults to None):
|
||||||
|
Number of selected groups for each token(for each token, ensuring the selected experts is only within `topk_group` groups).
|
||||||
|
num_experts_per_tok (`int`, *optional*, defaults to None):
|
||||||
|
Number of selected experts, None means dense model.
|
||||||
|
moe_layer_freq (`int`, *optional*, defaults to 1):
|
||||||
|
The frequency of the MoE layer: one expert layer for every `moe_layer_freq - 1` dense layers.
|
||||||
|
first_k_dense_replace (`int`, *optional*, defaults to 0):
|
||||||
|
Number of dense layers in shallow layers(embed->dense->dense->...->dense->moe->moe...->lm_head).
|
||||||
|
\--k dense layers--/
|
||||||
|
norm_topk_prob (`bool`, *optional*, defaults to False):
|
||||||
|
Whether to normalize the weights of the routed experts.
|
||||||
|
scoring_func (`str`, *optional*, defaults to 'softmax'):
|
||||||
|
Method of computing expert weights.
|
||||||
|
aux_loss_alpha (`float`, *optional*, defaults to 0.001):
|
||||||
|
Auxiliary loss weight coefficient.
|
||||||
|
seq_aux = (`bool`, *optional*, defaults to True):
|
||||||
|
Whether to compute the auxiliary loss for each individual sample.
|
||||||
|
num_key_value_heads (`int`, *optional*):
|
||||||
|
This is the number of key_value heads that should be used to implement Grouped Query Attention. If
|
||||||
|
`num_key_value_heads=num_attention_heads`, the model will use Multi Head Attention (MHA), if
|
||||||
|
`num_key_value_heads=1 the model will use Multi Query Attention (MQA) otherwise GQA is used. When
|
||||||
|
converting a multi-head checkpoint to a GQA checkpoint, each group key and value head should be constructed
|
||||||
|
by meanpooling all the original heads within that group. For more details checkout [this
|
||||||
|
paper](https://arxiv.org/pdf/2305.13245.pdf). If it is not specified, will default to
|
||||||
|
`num_attention_heads`.
|
||||||
|
hidden_act (`str` or `function`, *optional*, defaults to `"silu"`):
|
||||||
|
The non-linear activation function (function or string) in the decoder.
|
||||||
|
max_position_embeddings (`int`, *optional*, defaults to 2048):
|
||||||
|
The maximum sequence length that this model might ever be used with.
|
||||||
|
initializer_range (`float`, *optional*, defaults to 0.02):
|
||||||
|
The standard deviation of the truncated_normal_initializer for initializing all weight matrices.
|
||||||
|
rms_norm_eps (`float`, *optional*, defaults to 1e-06):
|
||||||
|
The epsilon used by the rms normalization layers.
|
||||||
|
use_cache (`bool`, *optional*, defaults to `True`):
|
||||||
|
Whether or not the model should return the last key/values attentions (not used by all models). Only
|
||||||
|
relevant if `config.is_decoder=True`.
|
||||||
|
pad_token_id (`int`, *optional*):
|
||||||
|
Padding token id.
|
||||||
|
bos_token_id (`int`, *optional*, defaults to 1):
|
||||||
|
Beginning of stream token id.
|
||||||
|
eos_token_id (`int`, *optional*, defaults to 2):
|
||||||
|
End of stream token id.
|
||||||
|
pretraining_tp (`int`, *optional*, defaults to 1):
|
||||||
|
Experimental feature. Tensor parallelism rank used during pretraining. Please refer to [this
|
||||||
|
document](https://huggingface.co/docs/transformers/parallelism) to understand more about it. This value is
|
||||||
|
necessary to ensure exact reproducibility of the pretraining results. Please refer to [this
|
||||||
|
issue](https://github.com/pytorch/pytorch/issues/76232).
|
||||||
|
tie_word_embeddings (`bool`, *optional*, defaults to `False`):
|
||||||
|
Whether to tie weight embeddings
|
||||||
|
rope_theta (`float`, *optional*, defaults to 10000.0):
|
||||||
|
The base period of the RoPE embeddings.
|
||||||
|
rope_scaling (`Dict`, *optional*):
|
||||||
|
Dictionary containing the scaling configuration for the RoPE embeddings. Currently supports two scaling
|
||||||
|
strategies: linear and dynamic. Their scaling factor must be a float greater than 1. The expected format is
|
||||||
|
`{"type": strategy name, "factor": scaling factor}`. When using this flag, don't update
|
||||||
|
`max_position_embeddings` to the expected new maximum.
|
||||||
|
attention_bias (`bool`, defaults to `False`, *optional*, defaults to `False`):
|
||||||
|
Whether to use a bias in the query, key, value and output projection layers during self-attention.
|
||||||
|
attention_dropout (`float`, *optional*, defaults to 0.0):
|
||||||
|
The dropout ratio for the attention probabilities.
|
||||||
|
|
||||||
|
```python
|
||||||
|
>>> from transformers import DeepseekV3Model, DeepseekV3Config
|
||||||
|
|
||||||
|
>>> # Initializing a Deepseek-V3 style configuration
|
||||||
|
>>> configuration = DeepseekV3Config()
|
||||||
|
|
||||||
|
>>> # Accessing the model configuration
|
||||||
|
>>> configuration = model.config
|
||||||
|
```"""
|
||||||
|
|
||||||
|
model_type = "deepseek_v3"
|
||||||
|
keys_to_ignore_at_inference = ["past_key_values"]
|
||||||
|
|
||||||
|
def __init__(
|
||||||
|
self,
|
||||||
|
vocab_size=129280,
|
||||||
|
hidden_size=7168,
|
||||||
|
intermediate_size=18432,
|
||||||
|
moe_intermediate_size=2048,
|
||||||
|
num_hidden_layers=61,
|
||||||
|
num_nextn_predict_layers=1,
|
||||||
|
num_attention_heads=128,
|
||||||
|
num_key_value_heads=128,
|
||||||
|
n_shared_experts=1,
|
||||||
|
n_routed_experts=256,
|
||||||
|
ep_size=1,
|
||||||
|
routed_scaling_factor=2.5,
|
||||||
|
kv_lora_rank=512,
|
||||||
|
q_lora_rank=1536,
|
||||||
|
qk_rope_head_dim=64,
|
||||||
|
v_head_dim=128,
|
||||||
|
qk_nope_head_dim=128,
|
||||||
|
topk_method="noaux_tc",
|
||||||
|
n_group=8,
|
||||||
|
topk_group=4,
|
||||||
|
num_experts_per_tok=8,
|
||||||
|
moe_layer_freq=1,
|
||||||
|
first_k_dense_replace=3,
|
||||||
|
norm_topk_prob=True,
|
||||||
|
scoring_func="sigmoid",
|
||||||
|
aux_loss_alpha=0.001,
|
||||||
|
seq_aux=True,
|
||||||
|
hidden_act="silu",
|
||||||
|
max_position_embeddings=4096,
|
||||||
|
initializer_range=0.02,
|
||||||
|
rms_norm_eps=1e-6,
|
||||||
|
use_cache=True,
|
||||||
|
pad_token_id=None,
|
||||||
|
bos_token_id=0,
|
||||||
|
eos_token_id=1,
|
||||||
|
pretraining_tp=1,
|
||||||
|
tie_word_embeddings=False,
|
||||||
|
rope_theta=10000.0,
|
||||||
|
rope_scaling=None,
|
||||||
|
attention_bias=False,
|
||||||
|
attention_dropout=0.0,
|
||||||
|
**kwargs,
|
||||||
|
):
|
||||||
|
self.vocab_size = vocab_size
|
||||||
|
self.max_position_embeddings = max_position_embeddings
|
||||||
|
self.hidden_size = hidden_size
|
||||||
|
self.intermediate_size = intermediate_size
|
||||||
|
self.moe_intermediate_size = moe_intermediate_size
|
||||||
|
self.num_hidden_layers = num_hidden_layers
|
||||||
|
self.num_nextn_predict_layers = num_nextn_predict_layers
|
||||||
|
self.num_attention_heads = num_attention_heads
|
||||||
|
self.n_shared_experts = n_shared_experts
|
||||||
|
self.n_routed_experts = n_routed_experts
|
||||||
|
self.ep_size = ep_size
|
||||||
|
self.routed_scaling_factor = routed_scaling_factor
|
||||||
|
self.kv_lora_rank = kv_lora_rank
|
||||||
|
self.q_lora_rank = q_lora_rank
|
||||||
|
self.qk_rope_head_dim = qk_rope_head_dim
|
||||||
|
self.v_head_dim = v_head_dim
|
||||||
|
self.qk_nope_head_dim = qk_nope_head_dim
|
||||||
|
self.topk_method = topk_method
|
||||||
|
self.n_group = n_group
|
||||||
|
self.topk_group = topk_group
|
||||||
|
self.num_experts_per_tok = num_experts_per_tok
|
||||||
|
self.moe_layer_freq = moe_layer_freq
|
||||||
|
self.first_k_dense_replace = first_k_dense_replace
|
||||||
|
self.norm_topk_prob = norm_topk_prob
|
||||||
|
self.scoring_func = scoring_func
|
||||||
|
self.aux_loss_alpha = aux_loss_alpha
|
||||||
|
self.seq_aux = seq_aux
|
||||||
|
# for backward compatibility
|
||||||
|
if num_key_value_heads is None:
|
||||||
|
num_key_value_heads = num_attention_heads
|
||||||
|
|
||||||
|
self.num_key_value_heads = num_key_value_heads
|
||||||
|
self.hidden_act = hidden_act
|
||||||
|
self.initializer_range = initializer_range
|
||||||
|
self.rms_norm_eps = rms_norm_eps
|
||||||
|
self.pretraining_tp = pretraining_tp
|
||||||
|
self.use_cache = use_cache
|
||||||
|
self.rope_theta = rope_theta
|
||||||
|
self.rope_scaling = rope_scaling
|
||||||
|
self.attention_bias = attention_bias
|
||||||
|
self.attention_dropout = attention_dropout
|
||||||
|
|
||||||
|
super().__init__(
|
||||||
|
pad_token_id=pad_token_id,
|
||||||
|
bos_token_id=bos_token_id,
|
||||||
|
eos_token_id=eos_token_id,
|
||||||
|
tie_word_embeddings=tie_word_embeddings,
|
||||||
|
**kwargs,
|
||||||
|
)
|
||||||
|
|
||||||
|
|
||||||
|
class MoonViTConfig(PretrainedConfig):
|
||||||
|
model_type = "moonvit"
|
||||||
|
|
||||||
|
def __init__(
|
||||||
|
self,
|
||||||
|
patch_size: int = 14,
|
||||||
|
init_pos_emb_height: int = 64,
|
||||||
|
init_pos_emb_width: int = 64,
|
||||||
|
num_attention_heads: int = 16,
|
||||||
|
num_hidden_layers: int = 27,
|
||||||
|
hidden_size: int = 1152,
|
||||||
|
intermediate_size: int = 4304,
|
||||||
|
merge_kernel_size: tuple[int, int] = (2, 2),
|
||||||
|
**kwargs,
|
||||||
|
):
|
||||||
|
super().__init__(**kwargs)
|
||||||
|
self.patch_size = patch_size
|
||||||
|
# Positional embedding config
|
||||||
|
self.init_pos_emb_height = init_pos_emb_height
|
||||||
|
self.init_pos_emb_width = init_pos_emb_width
|
||||||
|
# Transformer config
|
||||||
|
self.num_hidden_layers = num_hidden_layers
|
||||||
|
self.num_attention_heads = num_attention_heads
|
||||||
|
self.hidden_size = hidden_size
|
||||||
|
self.intermediate_size = intermediate_size
|
||||||
|
# Patch merger config
|
||||||
|
self.merge_kernel_size = merge_kernel_size
|
||||||
|
|
||||||
|
|
||||||
|
class KimiVLConfig(PretrainedConfig):
|
||||||
|
model_type = "kimi_vl"
|
||||||
|
|
||||||
|
def __init__(
|
||||||
|
self,
|
||||||
|
vision_config: Optional[Union[dict, MoonViTConfig]] = None,
|
||||||
|
text_config: Optional[Union[dict, DeepseekV3Config]] = None,
|
||||||
|
ignore_index: int = -100,
|
||||||
|
media_placeholder_token_id: int = 163605,
|
||||||
|
pad_token_id: int = 0,
|
||||||
|
**kwargs,
|
||||||
|
):
|
||||||
|
if vision_config is None:
|
||||||
|
vision_config = MoonViTConfig()
|
||||||
|
elif isinstance(vision_config, dict):
|
||||||
|
vision_config = MoonViTConfig(**vision_config)
|
||||||
|
self.vision_config = vision_config
|
||||||
|
|
||||||
|
if text_config is None:
|
||||||
|
text_config = DeepseekV3Config()
|
||||||
|
elif isinstance(text_config, dict):
|
||||||
|
text_config = DeepseekV3Config(**text_config)
|
||||||
|
self.text_config = text_config
|
||||||
|
|
||||||
|
self.ignore_index = ignore_index
|
||||||
|
self.media_placeholder_token_id = media_placeholder_token_id
|
||||||
|
|
||||||
|
attn_implementation = kwargs.get("attn_implementation")
|
||||||
|
if attn_implementation is not None:
|
||||||
|
if attn_implementation in ["eager", "flash_attention_2"]:
|
||||||
|
self._attn_implementation = attn_implementation
|
||||||
|
self.vision_config._attn_implementation = attn_implementation
|
||||||
|
self.text_config._attn_implementation = attn_implementation
|
||||||
|
else:
|
||||||
|
raise ValueError(
|
||||||
|
f"Invalid attention implementation: {attn_implementation}"
|
||||||
|
)
|
||||||
|
|
||||||
|
super().__init__(pad_token_id=pad_token_id, **kwargs)
|
||||||
BIN
figures/arch.png
Normal file
BIN
figures/arch.png
Normal file
Binary file not shown.
|
After Width: | Height: | Size: 626 KiB |
BIN
figures/demo1.png
Normal file
BIN
figures/demo1.png
Normal file
Binary file not shown.
|
After Width: | Height: | Size: 218 KiB |
BIN
figures/demo2.png
Normal file
BIN
figures/demo2.png
Normal file
Binary file not shown.
|
After Width: | Height: | Size: 258 KiB |
BIN
figures/logo.png
Normal file
BIN
figures/logo.png
Normal file
Binary file not shown.
|
After Width: | Height: | Size: 1.6 KiB |
BIN
figures/screenshot.png
Normal file
BIN
figures/screenshot.png
Normal file
Binary file not shown.
|
After Width: | Height: | Size: 808 KiB |
BIN
figures/thinking_perf.png
Normal file
BIN
figures/thinking_perf.png
Normal file
Binary file not shown.
|
After Width: | Height: | Size: 222 KiB |
9
generation_config.json
Normal file
9
generation_config.json
Normal file
@ -0,0 +1,9 @@
|
|||||||
|
{
|
||||||
|
"bos_token_id": 163584,
|
||||||
|
"pad_token_id": 163838,
|
||||||
|
"eos_token_id": [
|
||||||
|
163585
|
||||||
|
],
|
||||||
|
"do_sample": true,
|
||||||
|
"temperature": 0.6
|
||||||
|
}
|
||||||
126
image_processing_kimi_vl.py
Normal file
126
image_processing_kimi_vl.py
Normal file
@ -0,0 +1,126 @@
|
|||||||
|
"""Image processor class for KimiVL."""
|
||||||
|
|
||||||
|
import math
|
||||||
|
import numpy as np
|
||||||
|
from PIL import Image
|
||||||
|
from typing import Optional, Union
|
||||||
|
|
||||||
|
import torch
|
||||||
|
from torchvision.transforms import functional as TF
|
||||||
|
from transformers.image_utils import ImageInput, make_list_of_images, valid_images
|
||||||
|
from transformers.image_processing_utils import BaseImageProcessor, BatchFeature
|
||||||
|
from transformers.utils import TensorType
|
||||||
|
|
||||||
|
|
||||||
|
OPENAI_DATASET_MEAN = (0.48145466, 0.4578275, 0.40821073)
|
||||||
|
OPENAI_DATASET_STD = (0.26862954, 0.26130258, 0.27577711)
|
||||||
|
|
||||||
|
|
||||||
|
class KimiVLImageProcessor(BaseImageProcessor):
|
||||||
|
model_type = "kimi_vl"
|
||||||
|
|
||||||
|
def __init__(
|
||||||
|
self,
|
||||||
|
patch_size: int = 14,
|
||||||
|
pad_input: bool = False,
|
||||||
|
image_mean: tuple[float, float, float] = OPENAI_DATASET_MEAN,
|
||||||
|
image_std: tuple[float, float, float] = OPENAI_DATASET_STD,
|
||||||
|
in_token_limit: int = 4096,
|
||||||
|
merge_kernel_size: list[int, int] = [2, 2],
|
||||||
|
**kwargs,
|
||||||
|
):
|
||||||
|
super().__init__(**kwargs)
|
||||||
|
self.in_token_limit = in_token_limit
|
||||||
|
self.patch_size = patch_size
|
||||||
|
self.pad_input = pad_input
|
||||||
|
self.image_mean = image_mean
|
||||||
|
self.image_std = image_std
|
||||||
|
self.merge_kernel_size = merge_kernel_size
|
||||||
|
|
||||||
|
def rescale(
|
||||||
|
self, image: Image.Image, merge_kernel_size: list[int, int] = [2, 2]
|
||||||
|
) -> Image.Image:
|
||||||
|
w, h = image.size
|
||||||
|
patch_size = self.patch_size
|
||||||
|
|
||||||
|
if (w // patch_size) * (h // patch_size) > self.in_token_limit:
|
||||||
|
scale = math.sqrt(self.in_token_limit / ((w // patch_size) * (h // patch_size)))
|
||||||
|
new_w, new_h = int(w * scale), int(h * scale)
|
||||||
|
image = image.resize((new_w, new_h), Image.Resampling.BICUBIC)
|
||||||
|
if self.pad_input:
|
||||||
|
new_w, new_h = image.size
|
||||||
|
pad_size_h = merge_kernel_size[0] * patch_size
|
||||||
|
pad_size_w = merge_kernel_size[1] * patch_size
|
||||||
|
|
||||||
|
pad_h = (pad_size_h - new_h % pad_size_h) % pad_size_h
|
||||||
|
pad_w = (pad_size_w - new_w % pad_size_w) % pad_size_w
|
||||||
|
|
||||||
|
image = TF.pad(image, (0, 0, pad_w, pad_h))
|
||||||
|
else:
|
||||||
|
new_w, new_h = image.size
|
||||||
|
new_w = new_w - new_w % patch_size
|
||||||
|
new_h = new_h - new_h % patch_size
|
||||||
|
image = TF.center_crop(image, (new_h, new_w))
|
||||||
|
|
||||||
|
w, h = image.size
|
||||||
|
if w // patch_size >= 512 or h // patch_size >= 512:
|
||||||
|
raise ValueError("Exceed pos emb")
|
||||||
|
|
||||||
|
return image
|
||||||
|
|
||||||
|
def to_tensor(self, image: Image.Image) -> torch.Tensor:
|
||||||
|
return TF.to_tensor(image.convert("RGB"))
|
||||||
|
|
||||||
|
def normalize(self, image: torch.Tensor) -> torch.Tensor:
|
||||||
|
return TF.normalize(image, self.image_mean, self.image_std)
|
||||||
|
|
||||||
|
def patchify(self, image: torch.Tensor) -> tuple[torch.Tensor, list[int, int]]:
|
||||||
|
patch_size = self.patch_size
|
||||||
|
C, H, W = image.shape
|
||||||
|
patches = image.reshape(C, H // patch_size, patch_size, W // patch_size, patch_size)
|
||||||
|
patches = patches.permute(1, 3, 0, 2, 4)
|
||||||
|
patches = patches.contiguous().view(-1, C, patch_size, patch_size)
|
||||||
|
grid_hw = (H // patch_size, W // patch_size)
|
||||||
|
return patches, grid_hw
|
||||||
|
|
||||||
|
def _preprocess(self, image: ImageInput) -> tuple[torch.Tensor, list[int, int]]:
|
||||||
|
"""
|
||||||
|
Preprocess image and patchify it.
|
||||||
|
|
||||||
|
Args:
|
||||||
|
image (`ImageInput`):
|
||||||
|
Image to preprocess. Expects pixel values ranging from 0 to 255. If pixel values range from 0 to 1, set `do_rescale=False`.
|
||||||
|
|
||||||
|
Returns:
|
||||||
|
patches: torch.Tensor
|
||||||
|
grid_hw: list[int, int]
|
||||||
|
"""
|
||||||
|
image = self.rescale(image, self.merge_kernel_size)
|
||||||
|
image = self.to_tensor(image)
|
||||||
|
image = self.normalize(image)
|
||||||
|
patches, grid_hw = self.patchify(image)
|
||||||
|
return patches, grid_hw
|
||||||
|
|
||||||
|
def preprocess(
|
||||||
|
self,
|
||||||
|
images: ImageInput,
|
||||||
|
return_tensors: Optional[Union[str, TensorType]] = None,
|
||||||
|
) -> BatchFeature:
|
||||||
|
images = make_list_of_images(images)
|
||||||
|
|
||||||
|
if not valid_images(images):
|
||||||
|
raise ValueError(
|
||||||
|
"Invalid image type. Must be of type PIL.Image.Image, numpy.ndarray, "
|
||||||
|
"torch.Tensor, tf.Tensor or jax.ndarray."
|
||||||
|
)
|
||||||
|
|
||||||
|
pixel_values, image_grid_hws = [], []
|
||||||
|
for image in images:
|
||||||
|
patches, image_grid_hw = self._preprocess(image)
|
||||||
|
pixel_values.append(patches)
|
||||||
|
image_grid_hws.append(image_grid_hw)
|
||||||
|
pixel_values = torch.concat(pixel_values, dim=0)
|
||||||
|
image_grid_hws = np.array(image_grid_hws)
|
||||||
|
data = {"pixel_values": pixel_values, "image_grid_hws": image_grid_hws}
|
||||||
|
|
||||||
|
return BatchFeature(data=data, tensor_type=return_tensors)
|
||||||
3
model-00001-of-00007.safetensors
Normal file
3
model-00001-of-00007.safetensors
Normal file
@ -0,0 +1,3 @@
|
|||||||
|
version https://git-lfs.github.com/spec/v1
|
||||||
|
oid sha256:0fcc575e77d59bbf4504439266d770827a3018b56e2eecf4ab5a218c944d6326
|
||||||
|
size 135
|
||||||
3
model-00002-of-00007.safetensors
Normal file
3
model-00002-of-00007.safetensors
Normal file
@ -0,0 +1,3 @@
|
|||||||
|
version https://git-lfs.github.com/spec/v1
|
||||||
|
oid sha256:2b830e3948efd3257f27de6b2960c64a6704dafed27ebf6f53b9e1af3a859fe2
|
||||||
|
size 135
|
||||||
3
model-00003-of-00007.safetensors
Normal file
3
model-00003-of-00007.safetensors
Normal file
@ -0,0 +1,3 @@
|
|||||||
|
version https://git-lfs.github.com/spec/v1
|
||||||
|
oid sha256:729e85635b6f8c04a5a50e1330b6d190651f9a5901a34d3ee7caa0d20dbc6bbb
|
||||||
|
size 135
|
||||||
3
model-00004-of-00007.safetensors
Normal file
3
model-00004-of-00007.safetensors
Normal file
@ -0,0 +1,3 @@
|
|||||||
|
version https://git-lfs.github.com/spec/v1
|
||||||
|
oid sha256:2339e10a3d58cb9908660eb5b76ae75b3c0c091a773575083d4920a905ea7a45
|
||||||
|
size 135
|
||||||
3
model-00005-of-00007.safetensors
Normal file
3
model-00005-of-00007.safetensors
Normal file
@ -0,0 +1,3 @@
|
|||||||
|
version https://git-lfs.github.com/spec/v1
|
||||||
|
oid sha256:0dd4dd6c7179e98fb69402f7d4719c175ffcfa5f5cdea208fb7be9e2f6204230
|
||||||
|
size 135
|
||||||
3
model-00006-of-00007.safetensors
Normal file
3
model-00006-of-00007.safetensors
Normal file
@ -0,0 +1,3 @@
|
|||||||
|
version https://git-lfs.github.com/spec/v1
|
||||||
|
oid sha256:db8f22f385d5743cabc235eff38ac77ab7c77023f906e8c56b074b56cda67bb7
|
||||||
|
size 135
|
||||||
3
model-00007-of-00007.safetensors
Normal file
3
model-00007-of-00007.safetensors
Normal file
@ -0,0 +1,3 @@
|
|||||||
|
version https://git-lfs.github.com/spec/v1
|
||||||
|
oid sha256:3a4aac88787ee97d9a34627d0c03fc7883b98253a239fd93a5cc82a9c8f0d8a7
|
||||||
|
size 135
|
||||||
5686
model.safetensors.index.json
Normal file
5686
model.safetensors.index.json
Normal file
File diff suppressed because it is too large
Load Diff
2674
modeling_kimi_vl.py
Normal file
2674
modeling_kimi_vl.py
Normal file
File diff suppressed because it is too large
Load Diff
20
preprocessor_config.json
Normal file
20
preprocessor_config.json
Normal file
@ -0,0 +1,20 @@
|
|||||||
|
{
|
||||||
|
"auto_map": {
|
||||||
|
"AutoImageProcessor": "image_processing_kimi_vl.KimiVLImageProcessor",
|
||||||
|
"AutoProcessor": "processing_kimi_vl.KimiVLProcessor"
|
||||||
|
},
|
||||||
|
"in_token_limit": 16384,
|
||||||
|
"patch_size": 14,
|
||||||
|
"num_pooled_tokens": 1024,
|
||||||
|
"image_mean": [
|
||||||
|
0.5,
|
||||||
|
0.5,
|
||||||
|
0.5
|
||||||
|
],
|
||||||
|
"image_std": [
|
||||||
|
0.5,
|
||||||
|
0.5,
|
||||||
|
0.5
|
||||||
|
],
|
||||||
|
"pad_input": true
|
||||||
|
}
|
||||||
170
processing_kimi_vl.py
Normal file
170
processing_kimi_vl.py
Normal file
@ -0,0 +1,170 @@
|
|||||||
|
# coding=utf-8
|
||||||
|
# Copyright 2025 The Moonshot Team and HuggingFace Inc. team. All rights reserved.
|
||||||
|
#
|
||||||
|
# The code is based on the Qwen2VL processor (qwen2_vl/processing_qwen2_vl.py), but modified for KimiVL.
|
||||||
|
#
|
||||||
|
# Licensed under the Apache License, Version 2.0 (the "License");
|
||||||
|
# you may not use this file except in compliance with the License.
|
||||||
|
# You may obtain a copy of the License at
|
||||||
|
#
|
||||||
|
# http://www.apache.org/licenses/LICENSE-2.0
|
||||||
|
#
|
||||||
|
# Unless required by applicable law or agreed to in writing, software
|
||||||
|
# distributed under the License is distributed on an "AS IS" BASIS,
|
||||||
|
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
||||||
|
# See the License for the specific language governing permissions and
|
||||||
|
# limitations under the License.
|
||||||
|
"""
|
||||||
|
Processor class for KimiVL.
|
||||||
|
"""
|
||||||
|
|
||||||
|
from typing import List, Union
|
||||||
|
|
||||||
|
from transformers.feature_extraction_utils import BatchFeature
|
||||||
|
from transformers.image_utils import ImageInput
|
||||||
|
from transformers.processing_utils import ProcessingKwargs, ProcessorMixin, Unpack, _validate_images_text_input_order
|
||||||
|
from transformers.tokenization_utils_base import PreTokenizedInput, TextInput
|
||||||
|
from transformers.utils import logging
|
||||||
|
|
||||||
|
|
||||||
|
logger = logging.get_logger(__name__)
|
||||||
|
|
||||||
|
|
||||||
|
class KimiVLProcessorKwargs(ProcessingKwargs, total=False):
|
||||||
|
_defaults = {
|
||||||
|
"text_kwargs": {
|
||||||
|
"padding": False,
|
||||||
|
},
|
||||||
|
"images_kwargs": {},
|
||||||
|
}
|
||||||
|
|
||||||
|
|
||||||
|
class KimiVLProcessor(ProcessorMixin):
|
||||||
|
r"""
|
||||||
|
Constructs a KimiVL processor which wraps a KimiVL image processor and a tokenizer into a single processor.
|
||||||
|
|
||||||
|
[`KimiVLProcessor`] offers all the functionalities of [`KimiVLImageProcessor`] and [`TikTokenTokenizer`]. See the
|
||||||
|
[`~KimiVLProcessor.__call__`] and [`~KimiVLProcessor.decode`] for more information.
|
||||||
|
|
||||||
|
Args:
|
||||||
|
image_processor ([`KimiVLImageProcessor`], *optional*):
|
||||||
|
The image processor is a required input.
|
||||||
|
tokenizer ([`TikTokenTokenizer`], *optional*):
|
||||||
|
The tokenizer is a required input.
|
||||||
|
chat_template (`str`, *optional*): A Jinja template which will be used to convert lists of messages
|
||||||
|
in a chat into a tokenizable string.
|
||||||
|
"""
|
||||||
|
|
||||||
|
attributes = ["image_processor", "tokenizer"]
|
||||||
|
valid_kwargs = [ "chat_template"]
|
||||||
|
image_processor_class = "AutoImageProcessor"
|
||||||
|
tokenizer_class = "AutoTokenizer"
|
||||||
|
|
||||||
|
def __init__(
|
||||||
|
self,
|
||||||
|
image_processor=None,
|
||||||
|
tokenizer=None,
|
||||||
|
chat_template=None,
|
||||||
|
**kwargs,
|
||||||
|
):
|
||||||
|
self.image_token = "<|media_pad|>"
|
||||||
|
super().__init__(image_processor, tokenizer, chat_template=chat_template)
|
||||||
|
|
||||||
|
def __call__(
|
||||||
|
self,
|
||||||
|
images: ImageInput = None,
|
||||||
|
text: Union[TextInput, PreTokenizedInput, List[TextInput], List[PreTokenizedInput]] = None,
|
||||||
|
**kwargs: Unpack[KimiVLProcessorKwargs],
|
||||||
|
) -> BatchFeature:
|
||||||
|
"""
|
||||||
|
Main method to prepare for the model one or several sequences(s) and image(s). This method forwards the `text`
|
||||||
|
and `kwargs` arguments to TikTokenTokenizer's [`~TikTokenTokenizer.__call__`] if `text` is not `None` to encode
|
||||||
|
the text. To prepare the image(s), this method forwards the `images` and `kwrags` arguments to
|
||||||
|
CLIPImageProcessor's [`~CLIPImageProcessor.__call__`] if `images` is not `None`. Please refer to the docstring
|
||||||
|
of the above two methods for more information.
|
||||||
|
|
||||||
|
Args:
|
||||||
|
images (`PIL.Image.Image`, `np.ndarray`, `torch.Tensor`, `List[PIL.Image.Image]`, `List[np.ndarray]`, `List[torch.Tensor]`):
|
||||||
|
The image or batch of images to be prepared. Each image can be a PIL image, NumPy array or PyTorch
|
||||||
|
tensor. Both channels-first and channels-last formats are supported.
|
||||||
|
text (`str`, `List[str]`, `List[List[str]]`):
|
||||||
|
The sequence or batch of sequences to be encoded. Each sequence can be a string or a list of strings
|
||||||
|
(pretokenized string). If the sequences are provided as list of strings (pretokenized), you must set
|
||||||
|
`is_split_into_words=True` (to lift the ambiguity with a batch of sequences).
|
||||||
|
return_tensors (`str` or [`~utils.TensorType`], *optional*):
|
||||||
|
If set, will return tensors of a particular framework. Acceptable values are:
|
||||||
|
- `'tf'`: Return TensorFlow `tf.constant` objects.
|
||||||
|
- `'pt'`: Return PyTorch `torch.Tensor` objects.
|
||||||
|
- `'np'`: Return NumPy `np.ndarray` objects.
|
||||||
|
- `'jax'`: Return JAX `jnp.ndarray` objects.
|
||||||
|
|
||||||
|
Returns:
|
||||||
|
[`BatchFeature`]: A [`BatchFeature`] with the following fields:
|
||||||
|
|
||||||
|
- **input_ids** -- List of token ids to be fed to a model. Returned when `text` is not `None`.
|
||||||
|
- **attention_mask** -- List of indices specifying which tokens should be attended to by the model (when
|
||||||
|
`return_attention_mask=True` or if *"attention_mask"* is in `self.model_input_names` and if `text` is not
|
||||||
|
`None`).
|
||||||
|
- **pixel_values** -- Pixel values to be fed to a model. Returned when `images` is not `None`.
|
||||||
|
"""
|
||||||
|
if images is None and text is None:
|
||||||
|
raise ValueError("You have to specify at least one of `images` or `text`.")
|
||||||
|
|
||||||
|
# check if images and text inputs are reversed for BC
|
||||||
|
images, text = _validate_images_text_input_order(images, text)
|
||||||
|
|
||||||
|
output_kwargs = self._merge_kwargs(
|
||||||
|
KimiVLProcessorKwargs,
|
||||||
|
tokenizer_init_kwargs=self.tokenizer.init_kwargs,
|
||||||
|
**kwargs,
|
||||||
|
)
|
||||||
|
if images is not None:
|
||||||
|
image_inputs = self.image_processor(images, **output_kwargs["images_kwargs"])
|
||||||
|
image_grid_hws = image_inputs["image_grid_hws"]
|
||||||
|
else:
|
||||||
|
image_inputs = {}
|
||||||
|
image_grid_hws = None
|
||||||
|
|
||||||
|
if isinstance(text, str):
|
||||||
|
text = [text]
|
||||||
|
elif not isinstance(text, list) and not isinstance(text[0], str):
|
||||||
|
raise ValueError("Invalid input text. Please provide a string, or a list of strings")
|
||||||
|
|
||||||
|
if image_grid_hws is not None:
|
||||||
|
merge_length = self.image_processor.merge_kernel_size[0] * self.image_processor.merge_kernel_size[1]
|
||||||
|
index = 0
|
||||||
|
for i in range(len(text)):
|
||||||
|
while self.image_token in text[i]:
|
||||||
|
text[i] = text[i].replace(
|
||||||
|
self.image_token,
|
||||||
|
"<|placeholder|>" * (image_grid_hws[index].prod() // merge_length),
|
||||||
|
1,
|
||||||
|
)
|
||||||
|
index += 1
|
||||||
|
text[i] = text[i].replace("<|placeholder|>", self.image_token)
|
||||||
|
|
||||||
|
text_inputs = self.tokenizer(text, **output_kwargs["text_kwargs"])
|
||||||
|
return BatchFeature(data={**text_inputs, **image_inputs})
|
||||||
|
|
||||||
|
def batch_decode(self, *args, **kwargs):
|
||||||
|
"""
|
||||||
|
This method forwards all its arguments to LlamaTokenizerFast's [`~PreTrainedTokenizer.batch_decode`]. Please
|
||||||
|
refer to the docstring of this method for more information.
|
||||||
|
"""
|
||||||
|
return self.tokenizer.batch_decode(*args, **kwargs)
|
||||||
|
|
||||||
|
def decode(self, *args, **kwargs):
|
||||||
|
"""
|
||||||
|
This method forwards all its arguments to LlamaTokenizerFast's [`~PreTrainedTokenizer.decode`]. Please refer to
|
||||||
|
the docstring of this method for more information.
|
||||||
|
"""
|
||||||
|
return self.tokenizer.decode(*args, **kwargs)
|
||||||
|
|
||||||
|
@property
|
||||||
|
def model_input_names(self):
|
||||||
|
tokenizer_input_names = self.tokenizer.model_input_names
|
||||||
|
image_processor_input_names = self.image_processor.model_input_names
|
||||||
|
return list(dict.fromkeys(tokenizer_input_names + image_processor_input_names))
|
||||||
|
|
||||||
|
|
||||||
|
__all__ = ["KimiVLProcessorKwargs"]
|
||||||
3
tiktoken.model
Normal file
3
tiktoken.model
Normal file
@ -0,0 +1,3 @@
|
|||||||
|
version https://git-lfs.github.com/spec/v1
|
||||||
|
oid sha256:e065f594aa8ac9c7c583e31626a98bc910072b59f7c24fe9f5980653bdf936ae
|
||||||
|
size 132
|
||||||
302
tokenization_moonshot.py
Normal file
302
tokenization_moonshot.py
Normal file
@ -0,0 +1,302 @@
|
|||||||
|
import os
|
||||||
|
import tiktoken
|
||||||
|
|
||||||
|
from logging import getLogger
|
||||||
|
from pathlib import Path
|
||||||
|
from typing import (
|
||||||
|
cast,
|
||||||
|
Tuple,
|
||||||
|
Dict,
|
||||||
|
Iterator,
|
||||||
|
List,
|
||||||
|
Union,
|
||||||
|
Optional,
|
||||||
|
)
|
||||||
|
from shutil import copyfile
|
||||||
|
from tiktoken.load import load_tiktoken_bpe
|
||||||
|
from tokenizers import AddedToken
|
||||||
|
from transformers.tokenization_utils import PreTrainedTokenizer
|
||||||
|
from transformers.utils import to_py_obj
|
||||||
|
from transformers.models.gpt2.tokenization_gpt2 import bytes_to_unicode
|
||||||
|
|
||||||
|
|
||||||
|
logger = getLogger(__name__)
|
||||||
|
VOCAB_FILES_NAMES = {"vocab_file": "tiktoken.model"}
|
||||||
|
SPIECE_UNDERLINE = "▁"
|
||||||
|
|
||||||
|
|
||||||
|
class TikTokenTokenizer(PreTrainedTokenizer):
|
||||||
|
"""
|
||||||
|
Tokenizing and encoding/decoding text using the Tiktoken tokenizer. See megatron/tokenizer/tiktoken_tokenizer.py.
|
||||||
|
|
||||||
|
This tokenizer inherits from [`PreTrainedTokenizer`] which contains most of the main methods. Users should refer to
|
||||||
|
this superclass for more information regarding those methods.
|
||||||
|
|
||||||
|
Args:
|
||||||
|
vocab_file (`str`):
|
||||||
|
The path to the Tiktoken model file.
|
||||||
|
bos_token (`str` or `tokenizers.AddedToken`, *optional*, defaults to `"<|begin_of_text|>",`):
|
||||||
|
The beginning of sequence token that was used during pretraining. Can be used a sequence classifier token.
|
||||||
|
eos_token (`str` or `tokenizers.AddedToken`, *optional*, defaults to `"<|end_of_text|>"`):
|
||||||
|
The end of sequence token.
|
||||||
|
unk_token (`str` or `tokenizers.AddedToken`, *optional*, defaults to `"<|reserved_special_token_249|>"`):
|
||||||
|
The unknown token. A token that is not in the vocabulary cannot be converted to an ID and is set to be this
|
||||||
|
token instead. The second to last item in special_tokens.
|
||||||
|
pad_token (`str` or `tokenizers.AddedToken`, *optional*, defaults to `"<|reserved_special_token_250|>"`):
|
||||||
|
The token used for padding, for example when batching sequences of different lengths.
|
||||||
|
additional_special_tokens (list of `str`, *optional*):
|
||||||
|
A tuple or a list of additional tokens, which will be marked as `special`, meaning that they will be
|
||||||
|
skipped when decoding if `skip_special_tokens` is set to `True`.
|
||||||
|
"""
|
||||||
|
|
||||||
|
vocab_files_names = VOCAB_FILES_NAMES
|
||||||
|
|
||||||
|
model_input_names = ["input_ids", "attention_mask"]
|
||||||
|
|
||||||
|
special_tokens: Dict[str, int]
|
||||||
|
|
||||||
|
num_reserved_special_tokens = 256
|
||||||
|
|
||||||
|
pat_str = "|".join(
|
||||||
|
[
|
||||||
|
r"""[\p{Han}]+""",
|
||||||
|
r"""[^\r\n\p{L}\p{N}]?[\p{Lu}\p{Lt}\p{Lm}\p{Lo}\p{M}&&[^\p{Han}]]*[\p{Ll}\p{Lm}\p{Lo}\p{M}&&[^\p{Han}]]+(?i:'s|'t|'re|'ve|'m|'ll|'d)?""",
|
||||||
|
r"""[^\r\n\p{L}\p{N}]?[\p{Lu}\p{Lt}\p{Lm}\p{Lo}\p{M}&&[^\p{Han}]]+[\p{Ll}\p{Lm}\p{Lo}\p{M}&&[^\p{Han}]]*(?i:'s|'t|'re|'ve|'m|'ll|'d)?""",
|
||||||
|
r"""\p{N}{1,3}""",
|
||||||
|
r""" ?[^\s\p{L}\p{N}]+[\r\n]*""",
|
||||||
|
r"""\s*[\r\n]+""",
|
||||||
|
r"""\s+(?!\S)""",
|
||||||
|
r"""\s+""",
|
||||||
|
]
|
||||||
|
)
|
||||||
|
|
||||||
|
def __init__(
|
||||||
|
self,
|
||||||
|
vocab_file,
|
||||||
|
bos_token: Union[str, AddedToken] = "[BOS]",
|
||||||
|
eos_token: Union[str, AddedToken] = "[EOS]",
|
||||||
|
unk_token: Union[str, AddedToken] = "[UNK]",
|
||||||
|
pad_token: Union[str, AddedToken] = "[PAD]",
|
||||||
|
additional_special_tokens: Optional[List[str]] = None,
|
||||||
|
added_tokens_decoder: Optional[dict] = None,
|
||||||
|
**kwargs,
|
||||||
|
):
|
||||||
|
assert os.path.isfile(vocab_file), vocab_file
|
||||||
|
if additional_special_tokens is None:
|
||||||
|
additional_special_tokens = [
|
||||||
|
"<|im_end|>",
|
||||||
|
"<|im_middle|>",
|
||||||
|
"<|im_user|>",
|
||||||
|
"<|im_assistant|>",
|
||||||
|
"<|im_system|>",
|
||||||
|
]
|
||||||
|
special_tokens_mapping = {
|
||||||
|
i: added_tokens_decoder[i].content for i in added_tokens_decoder
|
||||||
|
}
|
||||||
|
|
||||||
|
self.vocab_file = vocab_file
|
||||||
|
mergeable_ranks = load_tiktoken_bpe(vocab_file)
|
||||||
|
num_base_tokens = len(mergeable_ranks)
|
||||||
|
self.special_tokens = {
|
||||||
|
special_tokens_mapping.get(i, f"<|reserved_token_{i}|>"): i
|
||||||
|
for i in range(
|
||||||
|
num_base_tokens, num_base_tokens + self.num_reserved_special_tokens + 2
|
||||||
|
)
|
||||||
|
}
|
||||||
|
|
||||||
|
self.model = tiktoken.Encoding(
|
||||||
|
name=Path(vocab_file).name,
|
||||||
|
pat_str=self.pat_str,
|
||||||
|
mergeable_ranks=mergeable_ranks,
|
||||||
|
special_tokens=self.special_tokens,
|
||||||
|
)
|
||||||
|
|
||||||
|
self.n_words: int = self.model.n_vocab
|
||||||
|
# BOS / EOS token IDs
|
||||||
|
self.bos_id: int = self.special_tokens[str(bos_token)]
|
||||||
|
self.eos_id: int = self.special_tokens[str(eos_token)]
|
||||||
|
|
||||||
|
self.pad_id: int = self.special_tokens[str(pad_token)]
|
||||||
|
self.unk_id: int = self.special_tokens[str(unk_token)]
|
||||||
|
|
||||||
|
self.byte_encoder = bytes_to_unicode()
|
||||||
|
self.byte_decoder = {v: k for k, v in self.byte_encoder.items()}
|
||||||
|
|
||||||
|
self.decoder = {}
|
||||||
|
for i in range(self.n_words):
|
||||||
|
# Taken from https://gist.github.com/xenova/a452a6474428de0182b17605a98631ee
|
||||||
|
decoding = "".join(
|
||||||
|
[
|
||||||
|
self.byte_encoder[ord(char)]
|
||||||
|
for char in self.model.decode_single_token_bytes(i).decode(
|
||||||
|
"latin-1"
|
||||||
|
)
|
||||||
|
]
|
||||||
|
)
|
||||||
|
self.decoder[i] = decoding
|
||||||
|
|
||||||
|
self.encoder = {}
|
||||||
|
for i in range(self.n_words):
|
||||||
|
if i in self.decoder:
|
||||||
|
self.encoder[self.decoder[i]] = i
|
||||||
|
|
||||||
|
super().__init__(
|
||||||
|
bos_token=bos_token,
|
||||||
|
eos_token=eos_token,
|
||||||
|
unk_token=unk_token,
|
||||||
|
pad_token=pad_token,
|
||||||
|
additional_special_tokens=additional_special_tokens,
|
||||||
|
**kwargs,
|
||||||
|
)
|
||||||
|
self.all_special_ids_set = set(self.all_special_ids)
|
||||||
|
|
||||||
|
def encode(
|
||||||
|
self, text: str, allow_special_tokens: bool = True, **kwargs
|
||||||
|
) -> List[int]:
|
||||||
|
"""
|
||||||
|
Encodes a string into a list of token IDs.
|
||||||
|
|
||||||
|
Args:
|
||||||
|
text (str): The input string to be encoded.
|
||||||
|
|
||||||
|
Returns:
|
||||||
|
list[int]: A list of token IDs.
|
||||||
|
"""
|
||||||
|
# If there are other args, we should call super().encode because there are a lot of code
|
||||||
|
# to handle those args. supper().encode finally will call _tokenize and _convert_token_to_id.
|
||||||
|
if len(kwargs) > 0:
|
||||||
|
return super().encode(text, **kwargs)
|
||||||
|
|
||||||
|
assert type(text) is str
|
||||||
|
|
||||||
|
# The tiktoken tokenizer can handle <=400k chars without
|
||||||
|
# pyo3_runtime.PanicException.
|
||||||
|
TIKTOKEN_MAX_ENCODE_CHARS = 400_000
|
||||||
|
|
||||||
|
# https://github.com/openai/tiktoken/issues/195
|
||||||
|
# Here we iterate over subsequences and split if we exceed the limit
|
||||||
|
# of max consecutive non-whitespace or whitespace characters.
|
||||||
|
MAX_NO_WHITESPACES_CHARS = 25_000
|
||||||
|
|
||||||
|
substrs = (
|
||||||
|
substr
|
||||||
|
for i in range(0, len(text), TIKTOKEN_MAX_ENCODE_CHARS)
|
||||||
|
for substr in self._split_whitespaces_or_nonwhitespaces(
|
||||||
|
text[i : i + TIKTOKEN_MAX_ENCODE_CHARS], MAX_NO_WHITESPACES_CHARS
|
||||||
|
)
|
||||||
|
)
|
||||||
|
t: List[int] = []
|
||||||
|
for substr in substrs:
|
||||||
|
if allow_special_tokens:
|
||||||
|
t.extend(
|
||||||
|
# we should consider special token as a common token
|
||||||
|
self.model.encode(
|
||||||
|
substr,
|
||||||
|
allowed_special="all",
|
||||||
|
)
|
||||||
|
)
|
||||||
|
else:
|
||||||
|
t.extend(
|
||||||
|
# we should consider special token as a common token
|
||||||
|
self.model.encode(
|
||||||
|
substr,
|
||||||
|
disallowed_special=(),
|
||||||
|
)
|
||||||
|
)
|
||||||
|
return t
|
||||||
|
|
||||||
|
def decode(self, token_ids: Union[int, List[int]], **kwargs) -> str:
|
||||||
|
"""
|
||||||
|
Decodes a list of token IDs into a string.
|
||||||
|
|
||||||
|
Args:
|
||||||
|
t (List[int]): The list of token IDs to be decoded.
|
||||||
|
|
||||||
|
Returns:
|
||||||
|
str: The decoded string.
|
||||||
|
"""
|
||||||
|
# If there are other args, we should call super().decode because there are a lot of code
|
||||||
|
# to handle those args. supper().encode finally will call convert_tokens_to_string and _convert_id_to_token.
|
||||||
|
if len(kwargs) > 0:
|
||||||
|
return super().decode(token_ids, **kwargs)
|
||||||
|
|
||||||
|
token_ids = to_py_obj(token_ids)
|
||||||
|
|
||||||
|
if type(token_ids) is int:
|
||||||
|
token_ids = [token_ids]
|
||||||
|
|
||||||
|
return self.model.decode(cast(List[int], token_ids))
|
||||||
|
|
||||||
|
@staticmethod
|
||||||
|
def _split_whitespaces_or_nonwhitespaces(
|
||||||
|
s: str, max_consecutive_slice_len: int
|
||||||
|
) -> Iterator[str]:
|
||||||
|
"""
|
||||||
|
Splits the string `s` so that each substring contains no more than `max_consecutive_slice_len`
|
||||||
|
consecutive whitespaces or consecutive non-whitespaces.
|
||||||
|
"""
|
||||||
|
current_slice_len = 0
|
||||||
|
current_slice_is_space = s[0].isspace() if len(s) > 0 else False
|
||||||
|
slice_start = 0
|
||||||
|
|
||||||
|
for i in range(len(s)):
|
||||||
|
is_now_space = s[i].isspace()
|
||||||
|
|
||||||
|
if current_slice_is_space ^ is_now_space:
|
||||||
|
current_slice_len = 1
|
||||||
|
current_slice_is_space = is_now_space
|
||||||
|
else:
|
||||||
|
current_slice_len += 1
|
||||||
|
if current_slice_len > max_consecutive_slice_len:
|
||||||
|
yield s[slice_start:i]
|
||||||
|
slice_start = i
|
||||||
|
current_slice_len = 1
|
||||||
|
yield s[slice_start:]
|
||||||
|
|
||||||
|
""" ----- Below are the abstract methods required by PreTrainedTokenizer ----- """
|
||||||
|
|
||||||
|
@property
|
||||||
|
def vocab_size(self) -> int:
|
||||||
|
return self.n_words
|
||||||
|
|
||||||
|
def get_vocab(self) -> Dict[str, int]:
|
||||||
|
return self.encoder
|
||||||
|
|
||||||
|
def _tokenize(self, text: str, **kwargs) -> List[str]:
|
||||||
|
return [self.decoder[t] for t in self.encode(text)]
|
||||||
|
|
||||||
|
def _convert_token_to_id(self, token: str) -> int:
|
||||||
|
return self.encoder.get(token, self.unk_id)
|
||||||
|
|
||||||
|
def _convert_id_to_token(self, index: int) -> str:
|
||||||
|
return self.decoder.get(index)
|
||||||
|
|
||||||
|
@staticmethod
|
||||||
|
def clean_up_tokenization(out_string: str) -> str:
|
||||||
|
return out_string
|
||||||
|
|
||||||
|
def convert_tokens_to_string(self, tokens: List[str]) -> str:
|
||||||
|
text = "".join(tokens).replace(SPIECE_UNDERLINE, "")
|
||||||
|
text = bytearray([self.byte_decoder[c] for c in text]).decode(
|
||||||
|
"utf-8", "replace"
|
||||||
|
)
|
||||||
|
return text
|
||||||
|
|
||||||
|
def save_vocabulary(
|
||||||
|
self, save_directory: str, filename_prefix: Optional[str] = None
|
||||||
|
) -> Tuple[str]:
|
||||||
|
if not os.path.isdir(save_directory):
|
||||||
|
logger.error(f"Vocabulary path ({save_directory}) should be a directory")
|
||||||
|
return
|
||||||
|
out_vocab_file = os.path.join(
|
||||||
|
save_directory,
|
||||||
|
(filename_prefix + "-" if filename_prefix else "")
|
||||||
|
+ VOCAB_FILES_NAMES["vocab_file"],
|
||||||
|
)
|
||||||
|
|
||||||
|
if os.path.abspath(self.vocab_file) != os.path.abspath(
|
||||||
|
out_vocab_file
|
||||||
|
) and os.path.isfile(self.vocab_file):
|
||||||
|
copyfile(self.vocab_file, out_vocab_file)
|
||||||
|
|
||||||
|
return (out_vocab_file,)
|
||||||
134
tokenizer_config.json
Normal file
134
tokenizer_config.json
Normal file
@ -0,0 +1,134 @@
|
|||||||
|
{
|
||||||
|
"added_tokens_decoder": {
|
||||||
|
"163584": {
|
||||||
|
"content": "[BOS]",
|
||||||
|
"lstrip": false,
|
||||||
|
"normalized": false,
|
||||||
|
"rstrip": false,
|
||||||
|
"single_word": false,
|
||||||
|
"special": true
|
||||||
|
},
|
||||||
|
"163585": {
|
||||||
|
"content": "[EOS]",
|
||||||
|
"lstrip": false,
|
||||||
|
"normalized": false,
|
||||||
|
"rstrip": false,
|
||||||
|
"single_word": false,
|
||||||
|
"special": true
|
||||||
|
},
|
||||||
|
"163586": {
|
||||||
|
"content": "<|im_end|>",
|
||||||
|
"lstrip": false,
|
||||||
|
"normalized": false,
|
||||||
|
"rstrip": false,
|
||||||
|
"single_word": false,
|
||||||
|
"special": true
|
||||||
|
},
|
||||||
|
"163601": {
|
||||||
|
"content": "<|im_middle|>",
|
||||||
|
"lstrip": false,
|
||||||
|
"normalized": false,
|
||||||
|
"rstrip": false,
|
||||||
|
"single_word": false,
|
||||||
|
"special": true
|
||||||
|
},
|
||||||
|
"163587": {
|
||||||
|
"content": "<|im_user|>",
|
||||||
|
"lstrip": false,
|
||||||
|
"normalized": false,
|
||||||
|
"rstrip": false,
|
||||||
|
"single_word": false,
|
||||||
|
"special": true
|
||||||
|
},
|
||||||
|
"163588": {
|
||||||
|
"content": "<|im_assistant|>",
|
||||||
|
"lstrip": false,
|
||||||
|
"normalized": false,
|
||||||
|
"rstrip": false,
|
||||||
|
"single_word": false,
|
||||||
|
"special": true
|
||||||
|
},
|
||||||
|
"163594": {
|
||||||
|
"content": "<|im_system|>",
|
||||||
|
"lstrip": false,
|
||||||
|
"normalized": false,
|
||||||
|
"rstrip": false,
|
||||||
|
"single_word": false,
|
||||||
|
"special": true
|
||||||
|
},
|
||||||
|
"163602": {
|
||||||
|
"content": "<|media_start|>",
|
||||||
|
"lstrip": false,
|
||||||
|
"normalized": false,
|
||||||
|
"rstrip": false,
|
||||||
|
"single_word": false,
|
||||||
|
"special": true
|
||||||
|
},
|
||||||
|
"163603": {
|
||||||
|
"content": "<|media_content|>",
|
||||||
|
"lstrip": false,
|
||||||
|
"normalized": false,
|
||||||
|
"rstrip": false,
|
||||||
|
"single_word": false,
|
||||||
|
"special": true
|
||||||
|
},
|
||||||
|
"163604": {
|
||||||
|
"content": "<|media_end|>",
|
||||||
|
"lstrip": false,
|
||||||
|
"normalized": false,
|
||||||
|
"rstrip": false,
|
||||||
|
"single_word": false,
|
||||||
|
"special": true
|
||||||
|
},
|
||||||
|
"163605": {
|
||||||
|
"content": "<|media_pad|>",
|
||||||
|
"lstrip": false,
|
||||||
|
"normalized": false,
|
||||||
|
"rstrip": false,
|
||||||
|
"single_word": false,
|
||||||
|
"special": true
|
||||||
|
},
|
||||||
|
"163838": {
|
||||||
|
"content": "[PAD]",
|
||||||
|
"lstrip": false,
|
||||||
|
"normalized": false,
|
||||||
|
"rstrip": false,
|
||||||
|
"single_word": false,
|
||||||
|
"special": true
|
||||||
|
},
|
||||||
|
"163839": {
|
||||||
|
"content": "[UNK]",
|
||||||
|
"lstrip": false,
|
||||||
|
"normalized": false,
|
||||||
|
"rstrip": false,
|
||||||
|
"single_word": false,
|
||||||
|
"special": true
|
||||||
|
}
|
||||||
|
},
|
||||||
|
"additional_special_tokens": [
|
||||||
|
"<|im_end|>",
|
||||||
|
"<|im_user|>",
|
||||||
|
"<|im_assistant|>",
|
||||||
|
"<|im_system|>",
|
||||||
|
"<|im_middle|>",
|
||||||
|
"<|media_start|>",
|
||||||
|
"<|media_content|>",
|
||||||
|
"<|media_end|>",
|
||||||
|
"<|media_pad|>"
|
||||||
|
],
|
||||||
|
"bos_token": "[BOS]",
|
||||||
|
"clean_up_tokenization_spaces": false,
|
||||||
|
"eos_token": "[EOS]",
|
||||||
|
"extra_special_tokens": {},
|
||||||
|
"model_max_length": 1048576,
|
||||||
|
"pad_token": "[PAD]",
|
||||||
|
"unk_token": "[UNK]",
|
||||||
|
"tokenizer_class": "TikTokenTokenizer",
|
||||||
|
"chat_template": "{%- for message in messages -%}{%- if loop.first and messages[0]['role'] != 'system' -%}{{'<|im_system|>system<|im_middle|>You are a helpful assistant<|im_end|>'}}{%- endif -%}{%- if message['role'] == 'system' -%}{{'<|im_system|>'}}{%- endif -%}{%- if message['role'] == 'user' -%}{{'<|im_user|>'}}{%- endif -%}{%- if message['role'] == 'assistant' -%}{{'<|im_assistant|>'}}{%- endif -%}{{- message['role'] -}}{{'<|im_middle|>'}}{%- if message['content'] is string -%}{{- message['content'] + '<|im_end|>' -}}{%- else -%}{%- for content in message['content'] -%}{%- if content['type'] == 'image' or 'image' in content or 'image_url' in content -%}{{'<|media_start|>image<|media_content|><|media_pad|><|media_end|>'}}{%- else -%}{{content['text']}}{%- endif -%}{%- endfor -%}{{'<|im_end|>'}}{%- endif -%}{%- endfor -%}{%- if add_generation_prompt -%}{{'<|im_assistant|>assistant<|im_middle|>'}}{%- endif -%}",
|
||||||
|
"auto_map": {
|
||||||
|
"AutoTokenizer": [
|
||||||
|
"tokenization_moonshot.TikTokenTokenizer",
|
||||||
|
null
|
||||||
|
]
|
||||||
|
}
|
||||||
|
}
|
||||||
Reference in New Issue
Block a user