deepj-tinystories

DeepJ TinyStories

An experimental 1,301,120-parameter language model trained with DeepJ. On 100 deterministic TinyStories validation windows (12800 tokens), loss was 2.182273 and perplexity was 8.866.

Model details

This compact language model was created with DeepJ and uses a DeepSeek-style Transformer architecture with 4 layers, a hidden size of 128, 4 attention heads, Q rank 64, KV rank 32 and a 2,048-token vocabulary.

It is not an exact implementation of DeepSeek V2, V3 or R1 and does not currently use an incremental KV cache. The model.dj checkpoint and tokenizer.bpe files use DeepJ's versioned binary formats. They are not directly loadable by the Python Transformers library.

Usage

Use DeepJ 0.8.0-alpha from Maven Central.

<dependency>
    <groupId>io.github.kirstenali</groupId>
    <artifactId>deepj</artifactId>
    <version>0.8.0-alpha</version>
</dependency>

Download this repository's model.dj and tokenizer.bpe into downloaded-model. Imports are omitted below.

Path directory = Path.of("downloaded-model");
BPEModel bpe = BPEModelIO.load(directory.resolve("tokenizer.bpe"));
BPETokenizer tokenizer = new BPETokenizer(bpe);
DeepJPrismConfig config = new DeepJPrismConfig(
        2048, 128, 128, 4, 4, 384, 64, 32, 0.2f, 1.0f);
DeepJPrism model = new DeepJPrism(config, 42L);
model.load(directory.resolve("model.dj"));
String text = TextGenerator.generate(model, tokenizer, config,
        "Once upon a time", 80, 0.8f, 40, 2026L);

Training data

Trained for 10000 steps with batch size 1 and 128-token sequences on TinyStories, a synthetic English dataset licensed CDLA-Sharing-1.0. The BPE vocabulary was trained from a bounded sample of the training split.

Limitations

Small experimental model trained only on synthetic children's stories. It may produce incorrect, repetitive, biased, or unsuitable text. It has not been evaluated for safety or downstream use and is not a general-purpose assistant.