PEFT
Safetensors
English
Chinese
long-context
context-extension
hierarchical-attention
segmented-attention
qwen3
lora
hici
Instructions to use ZengXiangyu/Qwen3-8b-HiCI-48k-500steps with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- PEFT
How to use ZengXiangyu/Qwen3-8b-HiCI-48k-500steps with PEFT:
from peft import PeftModel from transformers import AutoModelForCausalLM base_model = AutoModelForCausalLM.from_pretrained("./models/Qwen3-8B") model = PeftModel.from_pretrained(base_model, "ZengXiangyu/Qwen3-8b-HiCI-48k-500steps") - Notebooks
- Google Colab
- Kaggle
Download trainable_params.bin from ZengXiangyu/Qwen3-8b-HiCI-48k-500steps: direct link, hf CLI and curl.
- Browser
- Download file 4.29 GB
-
https://huggingface.co/ZengXiangyu/Qwen3-8b-HiCI-48k-500steps/resolve/main/trainable_params.bin
- Command line
-
hf download hf://ZengXiangyu/Qwen3-8b-HiCI-48k-500steps/trainable_params.bin
-
curl -L -o trainable_params.bin https://huggingface.co/ZengXiangyu/Qwen3-8b-HiCI-48k-500steps/resolve/main/trainable_params.bin
4.29 GB
- Xet hash:
- 4ca58deb8c279d9c39cad423feda0dcb2e52c31ddd0da049b13e9b3ee5eb07f7
- Size of remote file:
- 4.29 GB
- SHA256:
- 44f05270565489dee741fafc4c2230113332c5f34d57a33868590c27ac0dffb6
·
Xet efficiently stores Large Files inside Git, intelligently splitting files into unique chunks and accelerating uploads and downloads. More info.