yanyoyo
commited on
Commit
·
06a473e
1
Parent(s):
8515c91
update
Browse files
README.md
CHANGED
|
@@ -1,3 +1,51 @@
|
|
| 1 |
---
|
| 2 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 3 |
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
---
|
| 2 |
+
language:
|
| 3 |
+
- zh
|
| 4 |
+
- en
|
| 5 |
+
tags:
|
| 6 |
+
- internvl
|
| 7 |
+
- multimodal
|
| 8 |
+
- vision-language
|
| 9 |
+
- food
|
| 10 |
+
- finetuned
|
| 11 |
+
license: apache-2.0
|
| 12 |
+
datasets:
|
| 13 |
+
- food-recognition
|
| 14 |
+
model-index:
|
| 15 |
+
- name: InternVL2-2B-Food-Finetuned
|
| 16 |
+
results:
|
| 17 |
+
- task:
|
| 18 |
+
type: vision-language-understanding
|
| 19 |
+
name: food-recognition
|
| 20 |
---
|
| 21 |
+
|
| 22 |
+
# InternVL2-2B Food Recognition Finetuned Model
|
| 23 |
+
|
| 24 |
+
## Model Description
|
| 25 |
+
|
| 26 |
+
这是一个基于 InternVL2-2B 模型使用 LoRA 方法在食物识别数据集上微调的多模态模型。该模型专门优化了对食物图像的理解和描述能力。
|
| 27 |
+
|
| 28 |
+
### Key Features
|
| 29 |
+
|
| 30 |
+
- **基础模型**: InternVL2-2B
|
| 31 |
+
- **微调方法**: LoRA (Low-Rank Adaptation)
|
| 32 |
+
- **训练迭代**: 640 iterations
|
| 33 |
+
- **特定领域**: 食物识别与描述
|
| 34 |
+
- **多模态能力**: 图像理解和文本生成
|
| 35 |
+
|
| 36 |
+
## Training Details
|
| 37 |
+
|
| 38 |
+
### Base Model
|
| 39 |
+
- **架构**: InternVL2
|
| 40 |
+
- **参数量**: 2B
|
| 41 |
+
- **类型**: 视觉-语言多模态模型
|
| 42 |
+
|
| 43 |
+
### Fine-tuning
|
| 44 |
+
- **方法**: LoRA
|
| 45 |
+
- **配置文件**: internvl_v2_internlm2_2b_lora_finetune_food.py
|
| 46 |
+
- **训练步数**: 640
|
| 47 |
+
- **学习率**: 3.5e-5
|
| 48 |
+
- **训练轮数**: 10 epochs
|
| 49 |
+
|
| 50 |
+
## Usage
|
| 51 |
+
|