Add DeepSeek-V3.2 model metadata

[Copilot]

Adds model configuration for DeepSeek-V3.2, an enhanced 671B MoE model with 37B active parameters.

Changes

• Model metadata: Created models/deepseek/deepseek-v3.2/metadata.yaml
  • Runtime: vLLM >=0.6.5
  • Resources: 8x 80GB GPUs, 16 cores, 640GB RAM
  • Max context: 131K tokens
  • Sources: deepseek-ai/DeepSeek-V3.2 on HuggingFace and ModelScope

Configuration follows existing DeepSeek model patterns with resource requirements matching V3 baseline.

Original prompt

---

This section details on the original issue you should resolve

<issue_title>添加 deepseek-ai/DeepSeek-V3.2 模型</issue_title>
<issue_description></issue_description>

Comments on the Issue (you are @copilot in this section)

• Fixes 添加 deepseek-ai/DeepSeek-V3.2 模型 #60

---

💬 We'd love your input! Share your thoughts on Copilot coding agent in our 2 minute survey.