ERNIE 4.5 0.3B Base
The smallest ERNIE 4.5 language model uses a dense network. It provides a compact starting point for local text processing. This entry covers the base checkpoint.
Model design and access
This checkpoint ends at pretraining. It shares the small dense architecture with the chat release but has not received its instruction-focused post-training. It is supplied in Transformer-style PyTorch format.ERNIE-4.5-0.3B-Base-PT official model card
Practical use
Practical guidance. This small base model can make adaptation experiments accessible. Use held-out examples that differ from the training templates. Evaluate both task accuracy and whether the model can refuse malformed input after instruction tuning.ERNIE-4.5-0.3B-Base-PT official model card
Specifications & access
| Detail | Value and conditions |
|---|---|
| Access | Downloadable |
| Model identifiers | ERNIE-4.5-0.3B-Base-PT |
| Context window | 131,072 tokens Official checkpoint context window. Input and generation share this limit. ERNIE-4.5-0.3B-Base-PT official model card |
| Parameters | 0.36B dense parametersERNIE-4.5-0.3B-Base-PT official model card |
| Architecture | 18 layers; 16 query heads and two key/value headsERNIE-4.5-0.3B-Base-PT official model card |
| License | Apache 2.0ERNIE-4.5-0.3B-Base-PT official model card |
Related models
- ERNIE 4.5 0.3B
Baidu ยท ERNIE 4.5
Post-trained dense language model with 300 million parameters for local text processing.
Launch date not confirmedDownloadable