---
license: mit
language:
- en
- es
tags:
- small
- tiny
- tinyword
- theword
- harley-ml
- small-language-model
- word-generation
- word-generator
- text-generation
- qwen3
datasets:
- Harley-ml/es-en-words
---

# TinyWord-v2-128k

TinyWord-v2 is a revamped and retrained version of v1. In v1, we noticed that it didn't use weight-tying, which ate up half of its parameters. This was misleading as it was effectively the same size as MicroWord.
Anyway, this version achives much better performace compared to v1.

## Architecture

| Parameter | Value |
|---|---|
| Hidden Layers | 2 |
| Hidden Size | 48 |
| Attention Heads | 1 |
| KV Heads | 1 |
| Vocab Size | 1,200 |
| Intermediate Size | 160 |
| RoPE Theta | 1,000 |
| Max Position Embeddings | 32 |
| Tie Word Embeddings | True |

## Training

TinyWord-v2 was trained on 753,232 unique words (entries), 3,225,398 tokens, and 7,022,310 characters. ~660k of those words are English, while ~90k of them are Spanish.

### Dataset

|           Key           |   Value   |
| :---------------------: | :-------: |
|     Entries (words)     |  753,232  |
|          Tokens         | 3,225,398 |
|        Characters       | 7,022,310 |
|  Avg. Tokens Per Entry  |    ~4.2   |
|   Avg. Words Per Entry  |     1     |
|   Avg. Chars Per Entry  |    ~9.3   |
|  Longest Entry (Tokens) |     36    |
| Shortest Entry (Tokens) |     1     |
|      English Words      |   ~660k   |
|      Spanish Words      |    ~90k   |


### Hardware

TinyWord-v2 was trained on a NVIDA RTX 2060 6GB for 6 epochs with a batch size of 32.

### Training Results

| Step | Train Loss | Val Loss | Train PPL | Eval PPL |
|---|---|---|---|---|
| 2000 | 3.0579 | 2.5138 | 21.28 | 12.35 |
| 4000 | 2.0494 | 1.9456 | 7.76 | 6.99 |
| 6000 | 1.8572 | 1.7965 | 6.40 | 6.03 |
| 8000 | 1.7822 | 1.7294 | 5.94 | 5.64 |
| 10000 | 1.7360 | 1.6932 | 5.67 | 5.44 |

## Generations

Prompt: `w`

Output:
```
wrtervulatoration
```

Prompt: `app`

Output:
```
appatating
```

Prompt: `a`

Output:
```
ay's
```

Prompt: `z`

Output:
```
aceae
```

## Limitations

1. It does not generate sentences, prose, code, or anything besides a single word-like sequence.
2. It cannot reason or produce complex language.
3. Generated words may not be real. The goal isn't real word generation but reflecting the lexicon and morphology of the English and Spanish languages through tiny language models.
4. Output is non-deterministic. The same prompt can produce very different completions across runs.

# Inference

```python
# =============================================================================
# Inference
# =============================================================================

MODEL_DIR      = "Harley-ml/TinyWord2-128k"   # path
TOKENIZER_PATH = "Harley-ml/TinyWord2-128k"

# --- Generation settings ---
PROMPT             = "w"   # prompt
MAX_NEW_TOKENS     = 32
TEMPERATURE        = 1.2
TOP_P              = 0.95
TOP_K              = 50
REPETITION_PENALTY = 1.1
DO_SAMPLE          = True

# =============================================================================

import torch
from pathlib import Path
from transformers import (
    AutoModelForCausalLM,
    PreTrainedTokenizerFast,
    AddedToken,
)

# ---------------------------------------------------------------------------
# Device
# ---------------------------------------------------------------------------

device = (
    "cuda" if torch.cuda.is_available() else
    "mps"  if torch.backends.mps.is_available() else
    "cpu"
)
print(f"Device : {device}")

# ---------------------------------------------------------------------------
# Tokenizer  (mirrors training setup)
# ---------------------------------------------------------------------------

def load_tokenizer(path: str):
    p = Path(path).resolve()
    if not p.exists():
        raise FileNotFoundError(f"Tokenizer not found: {p}")
    tok = PreTrainedTokenizerFast(tokenizer_file=str(p))
    specials = {}
    if tok.bos_token is None: specials["bos_token"] = AddedToken("<|bos|>", special=True)
    if tok.eos_token is None: specials["eos_token"] = AddedToken("<|eos|>", special=True)
    if tok.unk_token is None: specials["unk_token"] = AddedToken("<|unk|>", special=True)
    if tok.pad_token is None:
        if tok.eos_token is not None:
            tok.pad_token = tok.eos_token
        else:
            specials["pad_token"] = AddedToken("<|pad|>", special=True)
    if specials:
        tok.add_special_tokens(specials)
    tok.padding_side = "left"  # left-pad for batched generation
    return tok

print("Loading tokenizer...")
tokenizer = load_tokenizer(TOKENIZER_PATH)
print(f"  Vocab size : {tokenizer.vocab_size}")
print(f"  BOS        : {tokenizer.bos_token!r}")
print(f"  EOS        : {tokenizer.eos_token!r}")
print(f"  PAD        : {tokenizer.pad_token!r}  (id={tokenizer.pad_token_id})")

# ---------------------------------------------------------------------------
# Model
# ---------------------------------------------------------------------------

print(f"\nLoading model from {MODEL_DIR} ...")
model = AutoModelForCausalLM.from_pretrained(
    MODEL_DIR,
    dtype=torch.float16 if device == "cuda" else torch.float32,
    low_cpu_mem_usage=True,
)
model.eval()
model.to(device)

total_params = sum(p.numel() for p in model.parameters())
print(f"  Parameters : {total_params:,}")

# ---------------------------------------------------------------------------
# Generation helper
# ---------------------------------------------------------------------------

def generate(
    prompt: str             = PROMPT,
    max_new_tokens: int     = MAX_NEW_TOKENS,
    temperature: float      = TEMPERATURE,
    top_p: float            = TOP_P,
    top_k: int              = TOP_K,
    repetition_penalty: float = REPETITION_PENALTY,
    do_sample: bool         = DO_SAMPLE,
) -> str:
    
    bos         = tokenizer.bos_token or ""
    full_prompt = bos + prompt

    inputs = tokenizer(
        full_prompt,
        return_tensors="pt",
        add_special_tokens=False,
    ).to(device)
    inputs.pop("token_type_ids", None)  # Qwen3 doesn't use this

    gen_kwargs = dict(
        max_new_tokens     = max_new_tokens,
        do_sample          = do_sample,
        repetition_penalty = repetition_penalty,
        eos_token_id       = tokenizer.eos_token_id,
        pad_token_id       = tokenizer.pad_token_id,
    )
    if do_sample:
        gen_kwargs["temperature"] = temperature
        gen_kwargs["top_p"]       = top_p
        gen_kwargs["top_k"]       = top_k

    with torch.inference_mode():
        output_ids = model.generate(**inputs, **gen_kwargs)

    # Strip the prompt tokens so we only return what was generated
    prompt_len = inputs["input_ids"].shape[-1]
    new_ids    = output_ids[0][prompt_len:]
    return tokenizer.decode(new_ids, skip_special_tokens=True)


# ---------------------------------------------------------------------------
# Run
# ---------------------------------------------------------------------------

if __name__ == "__main__":
    print(f"\nPrompt : {PROMPT!r}")
    print("-" * 60)

    output = generate(PROMPT)

    print("Generated:")
    print(output)
```

### Related Models

1. [PicoWord](https://huggingface.co/Harley-ml/PicoWord-5k)
2. [MicroWord](https://huggingface.co/Harley-ml/MicroWord-23k)
3. [TinyWord](https://huggingface.co/Harley-ml/TinyWord-134k)
4. [MediumWord](https://huggingface.co/Harley-ml/MediumWord-559k)
5. [LargeWord](https://huggingface.co/Harley-ml/LargeWord-1.5M)

## Citation

```bibtex
@misc{tinyword2-128k,
  title     = {TinyWord-134k: A Test of Morphological Compression in TLMs},
  author    = {Harley-ml},
  year      = {2026},
  url       = {https://huggingface.co/Harley-ml/TinyWord2-128k}
}
```