Unsupport Qwen3

[Montaro2017]

Do you plan to support Qwen3?

Exception in thread "main" java.lang.IllegalArgumentException: No enum constant com.github.tjake.jlama.model.ModelSupport.ModelType.QWEN3
	at java.base/java.lang.Enum.valueOf(Enum.java:293)
	at com.github.tjake.jlama.model.ModelSupport$ModelType.valueOf(ModelSupport.java:60)
	at com.github.tjake.jlama.safetensors.SafeTensorSupport.detectModel(SafeTensorSupport.java:109)
	at com.github.tjake.jlama.model.ModelSupport.loadModel(ModelSupport.java:180)
	at com.github.tjake.jlama.model.ModelSupport.loadModel(ModelSupport.java:129)
	at com.github.tjake.jlama.model.ModelSupport.loadModel(ModelSupport.java:88)

And can Chinese issues be resolved?

[koosty]

I have created #168 for this use case. But you need to implement it by yourself:
e.g.:

// assuming Qwen3 is compatible with Qwen2, otherwise implement custom model, config, and tokenizer for Qwen3
ModelSupport.ModelType modelType = new ModelSupport.ModelType() {
    @Override
    public Class<? extends AbstractModel> getModelClass() {
        return Qwen2Model.class;
    }
    @Override
    public Class<? extends Config> getConfigClass() {
        return Qwen2Config.class;
    }
    @Override
    public Class<? extends Tokenizer> getTokenizerClass() {
        return LlamaTokenizer.class;
    }
};

ModelSupport.register("QWEN3", modelType);
AbstractModel model = ModelSupport.loadModel(new File("Qwen_Qwen3-Embedding-0.6B"), DType.F32, DType.I8);