Files changed (1) hide show
  1. src/about.py +24 -63
src/about.py CHANGED
@@ -1,72 +1,33 @@
1
- from dataclasses import dataclass
2
- from enum import Enum
3
 
4
- @dataclass
5
- class Task:
6
- benchmark: str
7
- metric: str
8
- col_name: str
9
 
 
10
 
11
- # Select your tasks here
12
- # ---------------------------------------------------
13
- class Tasks(Enum):
14
- # task_key in the json file, metric_key in the json file, name to display in the leaderboard
15
- task0 = Task("anli_r1", "acc", "ANLI")
16
- task1 = Task("logiqa", "acc_norm", "LogiQA")
17
 
18
- NUM_FEWSHOT = 0 # Change with your few shot
19
- # ---------------------------------------------------
 
 
 
20
 
 
 
 
21
 
 
22
 
23
- # Your leaderboard name
24
- TITLE = """<h1 align="center" id="space-title">Demo leaderboard</h1>"""
25
 
26
- # What does your leaderboard evaluate?
27
- INTRODUCTION_TEXT = """
28
- Intro text
29
- """
30
 
31
- # Which evaluations are you running? how can people reproduce what you have?
32
- LLM_BENCHMARKS_TEXT = f"""
33
- ## How it works
34
-
35
- ## Reproducibility
36
- To reproduce our results, here is the commands you can run:
37
-
38
- """
39
-
40
- EVALUATION_QUEUE_TEXT = """
41
- ## Some good practices before submitting a model
42
-
43
- ### 1) Make sure you can load your model and tokenizer using AutoClasses:
44
- ```python
45
- from transformers import AutoConfig, AutoModel, AutoTokenizer
46
- config = AutoConfig.from_pretrained("your model name", revision=revision)
47
- model = AutoModel.from_pretrained("your model name", revision=revision)
48
- tokenizer = AutoTokenizer.from_pretrained("your model name", revision=revision)
49
- ```
50
- If this step fails, follow the error messages to debug your model before submitting it. It's likely your model has been improperly uploaded.
51
-
52
- Note: make sure your model is public!
53
- Note: if your model needs `use_remote_code=True`, we do not support this option yet but we are working on adding it, stay posted!
54
-
55
- ### 2) Convert your model weights to [safetensors](https://huggingface.co/docs/safetensors/index)
56
- It's a new format for storing weights which is safer and faster to load and use. It will also allow us to add the number of parameters of your model to the `Extended Viewer`!
57
-
58
- ### 3) Make sure your model has an open license!
59
- This is a leaderboard for Open LLMs, and we'd love for as many people as possible to know they can use your model 🤗
60
-
61
- ### 4) Fill up your model card
62
- When we add extra information about models to the leaderboard, it will be automatically taken from the model card
63
-
64
- ## In case of model failure
65
- If your model is displayed in the `FAILED` category, its execution stopped.
66
- Make sure you have followed the above steps first.
67
- If everything is done, check you can launch the EleutherAIHarness on your model locally, using the above command without modifications (you can add `--limit` to limit the number of examples per task).
68
- """
69
-
70
- CITATION_BUTTON_LABEL = "Copy the following snippet to cite these results"
71
- CITATION_BUTTON_TEXT = r"""
72
- """
 
1
+ # Report
 
2
 
3
+ ## 1. 模型及类别选择
 
 
 
 
4
 
5
+ 本次实验选用了三类大模型:Llama 3, Mistral 7B, ChatGPT。
6
 
7
+ - **Llama 3**:开源社区广泛使用,适合中英文任务。
8
+ - **Mistral 7B**:轻量级,适合边缘设备。
9
+ - **ChatGPT**:闭源,适合通用对话任务,表现最优。
 
 
 
10
 
11
+ | 模型名称 | 参数量 | 开源情况 | 主要用途 |
12
+ |------------|--------|---------|----------------|
13
+ | Llama 3 | 70B | 是 | 多语言任务 |
14
+ | Mistral 7B | 7B | 是 | 低功耗推理任务 |
15
+ | ChatGPT | 未公开 | 否 | 通用对话、推理任务 |
16
 
17
+ **选择理由**:
18
+ - Llama 3和Mistral为开源,方便定制与修改;
19
+ - ChatGPT性能优越,作为基准。
20
 
21
+ ---
22
 
23
+ ## 2. 系统实现细节
 
24
 
25
+ ### Gradio交互界面截图
26
+ ![Gradio界面](https://huggingface.co/spaces/15cgggggfh/work1/blob/main/interface.png)
 
 
27
 
28
+ ### 输入与输出流程图
29
+ ```mermaid
30
+ graph TD
31
+ A[用户输入] --> B[Gradio界面]
32
+ B --> C[模型推理]
33
+ C --> D[返回结果]