Downloads · 30 days
5
5% of all-time downloads
internvl2_5-26b-mpo-hf
jliu7350/internvl2_5-26b-mpo-hf
internvl2_5-26b-mpo-hf is a machine learning model from jliu7350. Use it for the machine learning task on the model card, and read the license before you ship it in a product. The card lists the license as mit.
Transformers-native (InternVLForConditionalGeneration) conversion of OpenGVLab/InternVL25-26B-MPO, produced with transformers' convertinternvlweightstohf.py. OpenGVLab publishes official -hf conversions only for the 2…
All-time downloads
107
Public
Parameters
25.5B
51 GB on disk
Likes
0
Public
Repo makeup
Click a slice to open those files.
.safetensors51 GB · 100%
At a glance
- License
- mit
- Model type
- internvl
- Access
- Public
- Created
- Sep 2, 2026
- Updated
- Sep 3, 2026
- SHA
- 1cf125b2
Base models
Model card
From the Hugging Face model README
InternVL2_5-26B-MPO-hf (community conversion)
Transformers-native (InternVLForConditionalGeneration) conversion of
OpenGVLab/InternVL2_5-26B-MPO,
produced with transformers' convert_internvl_weights_to_hf.py. OpenGVLab publishes
official -hf conversions only for the 2B and 8B sizes. Tokenizer files come from
OpenGVLab/InternVL2_5-8B-MPO-hf
(tokenizer.model is byte-identical across the 8B and 26B originals).
The conversion pipeline was validated on the 8B: converting
InternVL2_5-8B-MPO with the same script reproduces the official
InternVL2_5-8B-MPO-hf weights tensor-exactly, and a roundtrip back to the
original layout reproduces the source weights tensor-exactly.
License: MIT, inherited from the source model.
- Type
- internvl
- License
- mit
- Created
- Sep 2, 2026
- Updated
- Sep 3, 2026