Downloads · 30 days
0
Qwen3-8B-ISE-IHT
yone2426/Qwen3-8B-ISE-IHT
Qwen3-8B-ISE-IHT is a machine learning model from yone2426. Use it for the machine learning task on the model card, and read the license before you ship it in a product. It is set up for pytorch.
Four roles: system=0, user=1, data=2, assistant=3. This checkpoint includes the full fine-tuned backbone AND the learned ISE table.
Access
Public
Updated Sep 16, 2026
Repo size
16.4 GB
Likes
0
Public
Repo makeup
Click a slice to open those files.
.safetensors16.4 GB · 100%
At a glance
- Library
- pytorch
- Access
- Public
- Created
- Sep 16, 2026
- Updated
- Sep 16, 2026
- SHA
- 874c553d
Base models
Model card
From the Hugging Face model README
ISE IHT reproduction checkpoint
Four roles: system=0, user=1, data=2, assistant=3. This checkpoint includes the full fine-tuned backbone AND the learned ISE table.
Install the included package (pip install -e .), then use ise_repro.model.load_ise_checkpoint(path_or_repo) or the included generate.py. Loading only base_model/ with AutoModel omits ISE and is not a valid evaluation.
Derived from https://github.com/tongwu2020/ISE at d88eff2d34bc3b6ec03071091d7be8a0b94ba422. Qwen3 is an extension; the official implementation targets Llama. See training_metadata.json and data_manifest.json for exact settings and provenance. Base-model and source-data terms continue to apply.
- Library
- pytorch
- Created
- Sep 16, 2026
- Updated
- Sep 16, 2026