ERNIE 4.5 VL 424B A47B
The largest ERNIE 4.5 vision-language release provides a large sparse model for image and text understanding. This entry covers the post-trained checkpoint.
Model design and access
This chat checkpoint derives from ERNIE-4.5-VL-424B-A47B-Base. Its post-training includes supervised learning and reinforcement learning with verifiable rewards. Baidu publishes a 131,072-token window and PyTorch-format weights.ERNIE-4.5-VL-424B-A47B-PT official model card
Practical use
Practical guidance. Reserve evaluation capacity for documents or visual questions where smaller checkpoints make important errors. More total parameters do not establish better results on your image distribution. Compare correctness, serving capacity and preprocessing before selecting this scale.ERNIE-4.5-VL-424B-A47B-PT official model card
Specifications & access
| Detail | Value and conditions |
|---|---|
| Access | Downloadable |
| Model identifiers | ERNIE-4.5-VL-424B-A47B-PT |
| Context window | 131,072 tokens Official checkpoint context window. Input and generation share this limit. ERNIE-4.5-VL-424B-A47B-PT official model card |
| Parameters | 424B total parameters; 47B active per tokenERNIE-4.5-VL-424B-A47B-PT official model card |
| Architecture | 54 layers; separate text and vision expert poolsERNIE-4.5-VL-424B-A47B-PT official model card |
| License | Apache 2.0ERNIE-4.5-VL-424B-A47B-PT official model card |
Related models
- ERNIE 4.5 VL 424B A47B Base
Baidu ยท ERNIE 4.5
Large sparse vision-language base model with 424 billion total parameters and 47 billion active parameters.
Launch date not confirmedDownloadable