|
--- |
|
license: other |
|
license_name: glm-4 |
|
license_link: https://huggingface.co/THUDM/glm-4-9b/blob/main/LICENSE |
|
language: |
|
- zh |
|
pipeline_tag: text-generation |
|
--- |
|
# GLM-4-9b-chat Quantized with AutoFP8 |
|
|
|
使用 [m-a-p/COIG-CQIA](https://huggingface.co/datasets/m-a-p/COIG-CQIA) 的 COIG_pc 集作为校准量化的 [glm-4-9b-chat](https://huggingface.co/THUDM/glm-4-9b-chat) 模型。 |
|
|
|
主要为中文通常语言逻辑任务，为 vLLM 准备。 |
|
|
|
## 评估 |
|
|
|
使用 [lm-evaluation-harness](https://github.com/EleutherAI/lm-evaluation-harness/tree/7ad7c5b9d0f1c35c048af0ce8b197ebc2021dbd3) + vLLM 进行评估： |
|
|
|
|项目|THUDM/glm-4-9b-chat|此项目|Recovery| |
|
|---|---|---|---| |
|
|ceval-valid|71.84|70.36|97.94%| |
|
|cmmlu|72.23|70.42|97.49%| |
|
|agieval_logiqa_zh (5 shots)|44.24|39.32|88.88%| |
|
|**平均**|**62.77**|**60.03**|**95.63%**| |
|
|
