Add files using upload-large-folder tool
Browse files
README.md
CHANGED
|
@@ -1,49 +1,408 @@
|
|
| 1 |
---
|
| 2 |
-
|
|
|
|
|
|
|
|
|
|
| 3 |
library_name: transformers
|
| 4 |
-
tags:
|
| 5 |
-
- mergekit
|
| 6 |
-
- merge
|
| 7 |
-
|
| 8 |
---
|
| 9 |
-
|
| 10 |
-
|
| 11 |
-
|
| 12 |
-
|
| 13 |
-
|
| 14 |
-
|
| 15 |
-
|
| 16 |
-
|
| 17 |
-
|
| 18 |
-
|
| 19 |
-
|
| 20 |
-
|
| 21 |
-
|
| 22 |
-
|
| 23 |
-
|
| 24 |
-
|
| 25 |
-
|
| 26 |
-
|
| 27 |
-
|
| 28 |
-
|
| 29 |
-
|
| 30 |
-
|
| 31 |
-
|
| 32 |
-
|
| 33 |
-
|
| 34 |
-
|
| 35 |
-
|
| 36 |
-
|
| 37 |
-
|
| 38 |
-
|
| 39 |
-
|
| 40 |
-
|
| 41 |
-
|
| 42 |
-
|
| 43 |
-
|
| 44 |
-
|
| 45 |
-
|
| 46 |
-
|
| 47 |
-
|
| 48 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 49 |
```
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
---
|
| 2 |
+
language:
|
| 3 |
+
- zh
|
| 4 |
+
- en
|
| 5 |
+
pipeline_tag: text-generation
|
| 6 |
library_name: transformers
|
|
|
|
|
|
|
|
|
|
|
|
|
| 7 |
---
|
| 8 |
+
<div align="center">
|
| 9 |
+
<picture>
|
| 10 |
+
<img src="figures/joyai-logo.png" width="30%" alt="JoyAI-LLM Flash">
|
| 11 |
+
</picture>
|
| 12 |
+
</div>
|
| 13 |
+
<hr>
|
| 14 |
+
|
| 15 |
+
<div align="center" style="line-height: 1;">
|
| 16 |
+
<a href="https://huggingface.co/jdopensource" target="_blank"><img
|
| 17 |
+
alt="Hugging Face"
|
| 18 |
+
src="https://img.shields.io/badge/%F0%9F%A4%97%20Hugging%20Face-JD-ffc107?color=ffc107&logoColor=white"/></a>
|
| 19 |
+
<a
|
| 20 |
+
href="https://huggingface.co/jdopensource/JoyAI-LLM-Flash/blob/main/LICENSE"><img
|
| 21 |
+
alt="License"
|
| 22 |
+
src="https://img.shields.io/badge/License-Modified_MIT-f5de53?&color=f5de53"/></a>
|
| 23 |
+
</div>
|
| 24 |
+
|
| 25 |
+
|
| 26 |
+
|
| 27 |
+
|
| 28 |
+
## 1. Model Introduction
|
| 29 |
+
|
| 30 |
+
JoyAI-LLM-Flash is a state-of-the-art medium-sized instruct language model with
|
| 31 |
+
3 billion activated parameters and 48 billion total parameters. JoyAI-LLM-Flash
|
| 32 |
+
was pretrained on 20 trillion text tokens using Muon optimizer, followed by
|
| 33 |
+
large-scale supervised fine-tuning (SFT), direct preference optimization (DPO),
|
| 34 |
+
and reinforcement learning (RL) across diverse environments. JoyAI-LLM-Flash
|
| 35 |
+
achieves strong performance across frontier knowledge, reasoning, coding tasks
|
| 36 |
+
and agentic capabilities.
|
| 37 |
+
|
| 38 |
+
### Key Features
|
| 39 |
+
|
| 40 |
+
- Fiber Bundle RL: Introduces fiber bundle theory into reinforcement learning,
|
| 41 |
+
proposing a novel optimization framework, FiberPO. This method is
|
| 42 |
+
specifically designed to handle the challenges of large-scale and heterogeneous
|
| 43 |
+
agent training, improving stability and robustness under complex data
|
| 44 |
+
distributions.
|
| 45 |
+
- Training-Inference Collaboration: apply Muon optimizer with dense MTP,
|
| 46 |
+
develop novel optimization techniques to resolve instabilities while scaling
|
| 47 |
+
up, delivering 1.3× to 1.7× the throughput of the non-MTP version.
|
| 48 |
+
- Agentic Intelligence: designed for tool use, reasoning, and autonomous
|
| 49 |
+
problem-solving.
|
| 50 |
+
|
| 51 |
+
## 2. Model Summary
|
| 52 |
+
|
| 53 |
+
| | |
|
| 54 |
+
| :-----------------------------------------: | :----------------------: |
|
| 55 |
+
| **Architecture** | Mixture-of-Experts (MoE) |
|
| 56 |
+
| **Total Parameters** | 48B |
|
| 57 |
+
| **Activated Parameters** | 3B |
|
| 58 |
+
| **Number of Layers** (Dense layer included) | 40 |
|
| 59 |
+
| **Number of Dense Layers** | 1 |
|
| 60 |
+
| **Attention Hidden Dimension** | 2048 |
|
| 61 |
+
| **MoE Hidden Dimension** (per Expert) | 768 |
|
| 62 |
+
| **Number of Attention Heads** | 32 |
|
| 63 |
+
| **Number of Experts** | 256 |
|
| 64 |
+
| **Selected Experts per Token** | 8 |
|
| 65 |
+
| **Number of Shared Experts** | 1 |
|
| 66 |
+
| **Vocabulary Size** | 129K |
|
| 67 |
+
| **Context Length** | 128K |
|
| 68 |
+
| **Attention Mechanism** | MLA |
|
| 69 |
+
| **Activation Function** | SwiGLU |
|
| 70 |
+
| </div> | |
|
| 71 |
+
|
| 72 |
+
|
| 73 |
+
## 3. Evaluation Results
|
| 74 |
+
|
| 75 |
+
<table>
|
| 76 |
+
<thead>
|
| 77 |
+
<tr>
|
| 78 |
+
<th align="center">Benchmark</th>
|
| 79 |
+
<th align="center"><sup>JoyAI-LLM Flash</sup></th>
|
| 80 |
+
<th align="center"><sup>Qwen3-30B-A3B-Instuct-2507</sup></th>
|
| 81 |
+
<th align="center"><sup>GLM-4.7-Flash<br>(Non-thinking)</sup></th>
|
| 82 |
+
</tr>
|
| 83 |
+
</thead>
|
| 84 |
+
<tbody>
|
| 85 |
+
|
| 86 |
+
|
| 87 |
+
<tr>
|
| 88 |
+
<td align="center" colspan=8><strong>Knowledge & Alignment</strong></td>
|
| 89 |
+
</tr>
|
| 90 |
+
<tr>
|
| 91 |
+
<td align="center" style="vertical-align: middle">MMLU</td>
|
| 92 |
+
<td align="center" style="vertical-align: middle"><strong>89.50</strong></td>
|
| 93 |
+
<td align="center" style="vertical-align: middle">86.87</td>
|
| 94 |
+
<td align="center" style="vertical-align: middle">80.53</td>
|
| 95 |
+
</tr>
|
| 96 |
+
<tr>
|
| 97 |
+
<td align="center" style="vertical-align: middle">MMLU-Pro</td>
|
| 98 |
+
<td align="center" style="vertical-align: middle"><strong>81.02</strong></td>
|
| 99 |
+
<td align="center" style="vertical-align: middle">73.88</td>
|
| 100 |
+
<td align="center" style="vertical-align: middle">63.62</td>
|
| 101 |
+
</tr>
|
| 102 |
+
<tr>
|
| 103 |
+
<td align="center" style="vertical-align: middle">CMMLU</td>
|
| 104 |
+
<td align="center" style="vertical-align: middle"><strong>87.03</strong></td>
|
| 105 |
+
<td align="center" style="vertical-align: middle">85.88</td>
|
| 106 |
+
<td align="center" style="vertical-align: middle">75.85</td>
|
| 107 |
+
</tr>
|
| 108 |
+
<tr>
|
| 109 |
+
<td align="center" style="vertical-align: middle">GPQA-Diamond</td>
|
| 110 |
+
<td align="center" style="vertical-align: middle"><strong>74.43</strong></td>
|
| 111 |
+
<td align="center" style="vertical-align: middle">68.69</td>
|
| 112 |
+
<td align="center" style="vertical-align: middle">39.90</td>
|
| 113 |
+
</tr>
|
| 114 |
+
<tr>
|
| 115 |
+
<td align="center" style="vertical-align: middle">SuperGPQA</td>
|
| 116 |
+
<td align="center" style="vertical-align: middle"><strong>55.00</strong></td>
|
| 117 |
+
<td align="center" style="vertical-align: middle">52.00</td>
|
| 118 |
+
<td align="center" style="vertical-align: middle">32.00</td>
|
| 119 |
+
</tr>
|
| 120 |
+
<tr>
|
| 121 |
+
<td align="center" style="vertical-align: middle">LiveBench</td>
|
| 122 |
+
<td align="center" style="vertical-align: middle"><strong>72.90</strong></td>
|
| 123 |
+
<td align="center" style="vertical-align: middle">59.70</td>
|
| 124 |
+
<td align="center" style="vertical-align: middle">43.10</td>
|
| 125 |
+
</tr>
|
| 126 |
+
<tr>
|
| 127 |
+
<td align="center" style="vertical-align: middle">IFEval</td>
|
| 128 |
+
<td align="center" style="vertical-align: middle"><strong>86.69</strong></td>
|
| 129 |
+
<td align="center" style="vertical-align: middle">83.18</td>
|
| 130 |
+
<td align="center" style="vertical-align: middle">82.44</td>
|
| 131 |
+
</tr>
|
| 132 |
+
<tr>
|
| 133 |
+
<td align="center" style="vertical-align: middle">AlignBench</td>
|
| 134 |
+
<td align="center" style="vertical-align: middle"><strong>8.24</strong></td>
|
| 135 |
+
<td align="center" style="vertical-align: middle">8.07</td>
|
| 136 |
+
<td align="center" style="vertical-align: middle">6.85</td>
|
| 137 |
+
</tr>
|
| 138 |
+
<tr>
|
| 139 |
+
<td align="center" style="vertical-align: middle">HellaSwag</td>
|
| 140 |
+
<td align="center" style="vertical-align: middle"><strong>91.79</strong></td>
|
| 141 |
+
<td align="center" style="vertical-align: middle">89.90</td>
|
| 142 |
+
<td align="center" style="vertical-align: middle">60.84</td>
|
| 143 |
+
</tr>
|
| 144 |
+
|
| 145 |
+
<tr>
|
| 146 |
+
<td align="center" colspan=8><strong>Coding</strong></td>
|
| 147 |
+
</tr>
|
| 148 |
+
<tr>
|
| 149 |
+
<td align="center" style="vertical-align: middle">HumanEval</td>
|
| 150 |
+
<td align="center" style="vertical-align: middle"><strong>96.34</strong></td>
|
| 151 |
+
<td align="center" style="vertical-align: middle">95.12</td>
|
| 152 |
+
<td align="center" style="vertical-align: middle">74.39</td>
|
| 153 |
+
</tr>
|
| 154 |
+
<tr>
|
| 155 |
+
<td align="center" style="vertical-align: middle">LiveCodeBench</td>
|
| 156 |
+
<td align="center" style="vertical-align: middle"><strong>65.60</strong></td>
|
| 157 |
+
<td align="center" style="vertical-align: middle">39.71</td>
|
| 158 |
+
<td align="center" style="vertical-align: middle">27.43</td>
|
| 159 |
+
</tr>
|
| 160 |
+
<tr>
|
| 161 |
+
<td align="center" style="vertical-align: middle">SciCode</td>
|
| 162 |
+
<td align="center" style="vertical-align:
|
| 163 |
+
middle"><strong>3.08/22.92</strong></td>
|
| 164 |
+
<td align="center" style="vertical-align:
|
| 165 |
+
middle"><strong>3.08/22.92</strong></td>
|
| 166 |
+
<td align="center" style="vertical-align: middle">3.08/15.11</td>
|
| 167 |
+
</tr>
|
| 168 |
+
<tr>
|
| 169 |
+
<td align="center" colspan=8><strong>Mathematics</strong></td>
|
| 170 |
+
</tr>
|
| 171 |
+
<tr>
|
| 172 |
+
<td align="center" style="vertical-align: middle">GSM8K</td>
|
| 173 |
+
<td align="center" style="vertical-align: middle"><strong>95.83</strong></td>
|
| 174 |
+
<td align="center" style="vertical-align: middle">79.83</td>
|
| 175 |
+
<td align="center" style="vertical-align: middle">81.88</td>
|
| 176 |
+
</tr>
|
| 177 |
+
<tr>
|
| 178 |
+
<td align="center" style="vertical-align: middle">AIME2025</td>
|
| 179 |
+
<td align="center" style="vertical-align: middle"><strong>65.83</strong></td>
|
| 180 |
+
<td align="center" style="vertical-align: middle">62.08</td>
|
| 181 |
+
<td align="center" style="vertical-align: middle">24.17</td>
|
| 182 |
+
</tr>
|
| 183 |
+
<tr>
|
| 184 |
+
<td align="center" style="vertical-align: middle">MATH 500</td>
|
| 185 |
+
<td align="center" style="vertical-align: middle"><strong>97.10</strong></td>
|
| 186 |
+
<td align="center" style="vertical-align: middle">89.80</td>
|
| 187 |
+
<td align="center" style="vertical-align: middle">90.90</td>
|
| 188 |
+
</tr>
|
| 189 |
+
|
| 190 |
+
<tr>
|
| 191 |
+
<td align="center" colspan=8><strong>Agentic</strong></td>
|
| 192 |
+
</tr>
|
| 193 |
+
<tr>
|
| 194 |
+
<td align="center" style="vertical-align: middle">SWE-bench Verified</td>
|
| 195 |
+
<td align="center" style="vertical-align: middle"><strong>60.60</strong></td>
|
| 196 |
+
<td align="center" style="vertical-align: middle">24.44</td>
|
| 197 |
+
<td align="center" style="vertical-align: middle">51.60</td>
|
| 198 |
+
</tr>
|
| 199 |
+
<tr>
|
| 200 |
+
<td align="center" style="vertical-align: middle">Tau2-Retail</td>
|
| 201 |
+
<td align="center" style="vertical-align: middle"><strong>67.55</strong></td>
|
| 202 |
+
<td align="center" style="vertical-align: middle">53.51</td>
|
| 203 |
+
<td align="center" style="vertical-align: middle">62.28</td>
|
| 204 |
+
</tr>
|
| 205 |
+
<tr>
|
| 206 |
+
<td align="center" style="vertical-align: middle">Tau2-Airline</td>
|
| 207 |
+
<td align="center" style="vertical-align: middle"><strong>54.00</strong></td>
|
| 208 |
+
<td align="center" style="vertical-align: middle">32.00</td>
|
| 209 |
+
<td align="center" style="vertical-align: middle">52.00</td>
|
| 210 |
+
</tr>
|
| 211 |
+
<tr>
|
| 212 |
+
<td align="center" style="vertical-align: middle">Tau2-Telecom</td>
|
| 213 |
+
<td align="center" style="vertical-align: middle">79.83</td>
|
| 214 |
+
<td align="center" style="vertical-align: middle">4.39</td>
|
| 215 |
+
<td align="center" style="vertical-align: middle"><strong>88.60</strong></td>
|
| 216 |
+
</tr>
|
| 217 |
+
|
| 218 |
+
<tr>
|
| 219 |
+
<td align="center" colspan=8><strong>Long Context</strong></td>
|
| 220 |
+
</tr>
|
| 221 |
+
<tr>
|
| 222 |
+
<td align="center" style="vertical-align: middle">RULER</td>
|
| 223 |
+
<td align="center" style="vertical-align: middle"><strong>95.60</strong></td>
|
| 224 |
+
<td align="center" style="vertical-align: middle">89.66</td>
|
| 225 |
+
<td align="center" style="vertical-align: middle">56.12</td>
|
| 226 |
+
</tr>
|
| 227 |
+
</tbody>
|
| 228 |
+
</table>
|
| 229 |
+
|
| 230 |
+
|
| 231 |
+
## 4. Deployment
|
| 232 |
+
|
| 233 |
+
> [!Note]
|
| 234 |
+
> You can access JoyAI-LLM Flash API on https://docs.jdcloud.com/cn/jdaip/chat
|
| 235 |
+
> and we provide OpenAI/Anthropic-compatible API for you.
|
| 236 |
+
> Currently, JoyAI-LLM-Flash-Block-INT8 is recommended to run on the following
|
| 237 |
+
> inference engines:
|
| 238 |
+
|
| 239 |
+
* SGLang
|
| 240 |
+
|
| 241 |
+
Deployment examples can be found in the [Model Deployment
|
| 242 |
+
Guide](docs/deploy_guidance.md).
|
| 243 |
+
|
| 244 |
+
|
| 245 |
+
|
| 246 |
+
## 5. Model Usage
|
| 247 |
+
|
| 248 |
+
The usage demos below demonstrate how to call our official API.
|
| 249 |
+
|
| 250 |
+
For third-party APIs deployed with vLLM or SGLang, please note that:
|
| 251 |
+
|
| 252 |
+
> [!Note] Recommended sampling parameters: `temperature=0.6`, `top_p=1.0`
|
| 253 |
+
|
| 254 |
+
### Chat Completion
|
| 255 |
+
|
| 256 |
+
This is a simple chat completion script which shows how to call JoyAI-Flash
|
| 257 |
+
API.
|
| 258 |
+
|
| 259 |
+
```python
|
| 260 |
+
from openai import OpenAI
|
| 261 |
+
|
| 262 |
+
client = OpenAI(base_url="http://IP:PORT/v1", api_key="EMPTY")
|
| 263 |
+
|
| 264 |
+
|
| 265 |
+
def simple_chat(client: OpenAI):
|
| 266 |
+
messages = [
|
| 267 |
+
{
|
| 268 |
+
"role": "user",
|
| 269 |
+
"content": [
|
| 270 |
+
{
|
| 271 |
+
"type": "text",
|
| 272 |
+
"text": "which one is bigger, 9.11 or 9.9? think
|
| 273 |
+
carefully.",
|
| 274 |
+
}
|
| 275 |
+
],
|
| 276 |
+
},
|
| 277 |
+
]
|
| 278 |
+
model_name = client.models.list().data[0].id
|
| 279 |
+
response = client.chat.completions.create(
|
| 280 |
+
model=model_name, messages=messages, stream=False, max_tokens=4096
|
| 281 |
+
)
|
| 282 |
+
print(f"response: {response.choices[0].message.content}")
|
| 283 |
+
|
| 284 |
+
|
| 285 |
+
if __name__ == "__main__":
|
| 286 |
+
simple_chat(client)
|
| 287 |
```
|
| 288 |
+
|
| 289 |
+
|
| 290 |
+
### Tool call Completion
|
| 291 |
+
|
| 292 |
+
This is a simple toll call completion script which shows how to call
|
| 293 |
+
JoyAI-Flash API.
|
| 294 |
+
|
| 295 |
+
```python
|
| 296 |
+
import json
|
| 297 |
+
|
| 298 |
+
from openai import OpenAI
|
| 299 |
+
|
| 300 |
+
client = OpenAI(base_url="http://IP:PORT/v1", api_key="EMPTY")
|
| 301 |
+
|
| 302 |
+
|
| 303 |
+
def my_calculator(expression: str) -> str:
|
| 304 |
+
return str(eval(expression))
|
| 305 |
+
|
| 306 |
+
|
| 307 |
+
def rewrite(expression: str) -> str:
|
| 308 |
+
return str(expression)
|
| 309 |
+
|
| 310 |
+
|
| 311 |
+
def simple_tool_call(client: OpenAI):
|
| 312 |
+
messages = [
|
| 313 |
+
{
|
| 314 |
+
"role": "user",
|
| 315 |
+
"content": [
|
| 316 |
+
{
|
| 317 |
+
"type": "text",
|
| 318 |
+
"text": "use my functions to compute the results for the
|
| 319 |
+
equations: 6+1",
|
| 320 |
+
},
|
| 321 |
+
],
|
| 322 |
+
},
|
| 323 |
+
]
|
| 324 |
+
tools = [
|
| 325 |
+
{
|
| 326 |
+
"type": "function",
|
| 327 |
+
"function": {
|
| 328 |
+
"name": "my_calculator",
|
| 329 |
+
"description": "A calculator that can evaluate a mathematical
|
| 330 |
+
equation and compute its results.",
|
| 331 |
+
"parameters": {
|
| 332 |
+
"type": "object",
|
| 333 |
+
"properties": {
|
| 334 |
+
"expression": {
|
| 335 |
+
"type": "string",
|
| 336 |
+
"description": "The mathematical expression to
|
| 337 |
+
evaluate.",
|
| 338 |
+
},
|
| 339 |
+
},
|
| 340 |
+
"required": ["expression"],
|
| 341 |
+
},
|
| 342 |
+
},
|
| 343 |
+
},
|
| 344 |
+
{
|
| 345 |
+
"type": "function",
|
| 346 |
+
"function": {
|
| 347 |
+
"name": "rewrite",
|
| 348 |
+
"description": "Rewrite a given text for improved clarity",
|
| 349 |
+
"parameters": {
|
| 350 |
+
"type": "object",
|
| 351 |
+
"properties": {
|
| 352 |
+
"text": {
|
| 353 |
+
"type": "string",
|
| 354 |
+
"description": "The input text to rewrite",
|
| 355 |
+
}
|
| 356 |
+
},
|
| 357 |
+
},
|
| 358 |
+
},
|
| 359 |
+
},
|
| 360 |
+
]
|
| 361 |
+
model_name = client.models.list().data[0].id
|
| 362 |
+
response = client.chat.completions.create(
|
| 363 |
+
model=model_name,
|
| 364 |
+
messages=messages,
|
| 365 |
+
temperature=1.0,
|
| 366 |
+
max_tokens=1024,
|
| 367 |
+
tools=tools,
|
| 368 |
+
tool_choice="auto",
|
| 369 |
+
)
|
| 370 |
+
tool_calls = response.choices[0].message.tool_calls
|
| 371 |
+
|
| 372 |
+
results = []
|
| 373 |
+
for tool_call in tool_calls:
|
| 374 |
+
function_name = tool_call.function.name
|
| 375 |
+
function_args = tool_call.function.arguments
|
| 376 |
+
if function_name == "my_calculator":
|
| 377 |
+
result = my_calculator(**json.loads(function_args))
|
| 378 |
+
results.append(result)
|
| 379 |
+
messages.append({"role": "assistant", "tool_calls": tool_calls})
|
| 380 |
+
for tool_call, result in zip(tool_calls, results):
|
| 381 |
+
messages.append(
|
| 382 |
+
{
|
| 383 |
+
"role": "tool",
|
| 384 |
+
"tool_call_id": tool_call.id,
|
| 385 |
+
"name": tool_call.function.name,
|
| 386 |
+
"content": result,
|
| 387 |
+
}
|
| 388 |
+
)
|
| 389 |
+
response = client.chat.completions.create(
|
| 390 |
+
model=model_name,
|
| 391 |
+
messages=messages,
|
| 392 |
+
temperature=1.0,
|
| 393 |
+
max_tokens=1024,
|
| 394 |
+
)
|
| 395 |
+
print(response.choices[0].message.content)
|
| 396 |
+
|
| 397 |
+
|
| 398 |
+
if __name__ == "__main__":
|
| 399 |
+
simple_tool_call(client)
|
| 400 |
+
|
| 401 |
+
```
|
| 402 |
+
|
| 403 |
+
---
|
| 404 |
+
|
| 405 |
+
## 6. License
|
| 406 |
+
|
| 407 |
+
Both the code repository and the model weights are released under the [Modified
|
| 408 |
+
MIT License](LICENSE).
|