Downloads · 30 days
0
qwen3vl_8b_w4a8_lmhead4bit
gordon-zou/qwen3vl_8b_w4a8_lmhead4bit
qwen3vl_8b_w4a8_lmhead4bit is a text generation model from gordon-zou. Use it when you need the model to write or continue text.
--- license: apache-2.0 language: - en - zh basemodel: - Qwen/Qwen3-VL-8B-Instruct ---When running Qwen-image2.1, the lmhead doesn’t participate in computation, so I reduced it to 4-bit. Compared to the official W4A8…
Access
Public
Updated Sep 23, 2026
Repo size
6 GB
Likes
1
Public
Repo makeup
Click a slice to open those files.
.safetensors6 GB · 100%
At a glance
- Task
- Text Generation
- Access
- Public
- Created
- Sep 23, 2026
- Updated
- Sep 23, 2026
- SHA
- ab326dba
Try a prompt
Model card
From the Hugging Face model README
license: apache-2.0 language:
- en
- zh base_model:
- Qwen/Qwen3-VL-8B-Instruct ---When running Qwen-image2.1, the lm_head doesn’t participate in computation, so I reduced it to 4-bit. Compared to the official W4A8 version, the actual calculation accuracy didn’t change at all, and the size went down by 0.28 GB.
- Task
- Text Generation
- Created
- Sep 23, 2026
- Updated
- Sep 23, 2026