Instructions to use tampajohn/meow-lite-v5 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use tampajohn/meow-lite-v5 with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="tampajohn/meow-lite-v5")

# pip install -U transformers accelerate
# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("tampajohn/meow-lite-v5")
model = AutoModelForCausalLM.from_pretrained("tampajohn/meow-lite-v5", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use tampajohn/meow-lite-v5 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "tampajohn/meow-lite-v5"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "tampajohn/meow-lite-v5",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/tampajohn/meow-lite-v5

SGLang

How to use tampajohn/meow-lite-v5 with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "tampajohn/meow-lite-v5" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "tampajohn/meow-lite-v5",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "tampajohn/meow-lite-v5" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "tampajohn/meow-lite-v5",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use tampajohn/meow-lite-v5 with Docker Model Runner:
```
docker model run hf.co/tampajohn/meow-lite-v5
```

meow-lite-v5

The meow-lite family

meow-lite — v4 classic: the 104K meow toy behind the OpenAI/Anthropic shim (497 organic downloads and counting)
meow-lite-v5 — the chaotic cat: 6.8M from-scratch BPE GPT, 67.7% held-out comprehension, 3% leak
meow-lite-v6 — the calmer cat: same architecture, 70.4% comprehension, 3% leak
meow-lite-v6-dataset — the 16,633-pair teacher web + full methodology
github.com/tampajohn/meow-lite — server, specs, eval batteries (MIT) The chaotic cat. A 6,836,224-parameter from-scratch BPE GPT that reads English and can only speak cat. Maximum comprehension, minimal manners. It responds to everything a cat would: correctly, and then some.

Two temperaments

	v5 (this)	v6
Temperament	chaotic	calmer
Held-out synonym accuracy	67.7%	70.4%
Fresh-neutral leak	3%	3%
Vibe	comprehends more, interrupts more	comprehends slightly less, minds its manners

Served with action damping d=2 (inference-time prior suppression): the numbers above are the served reality, measured on the full held-out battery. Both were trained on the identical corpus with the identical recipe; the difference is which seed the dice favored. We ship both, because cats are inconsistent.

Architecture

GPT-2 (from-scratch, random init): n_layer=6, n_head=8, n_embd=256, n_positions=192
Tokenizer: byte-level BPE (~8k) + 10 cat action tokens as specials
Trained in minutes on a laptop (the 16,633-pair teacher web took ~13h on 2x GB10)
Output hard-masked to cat tokens (CatMask LogitsProcessor): non-cat output is impossible by construction
ActionOnce: each action token fires once per response (repeat biting 100% -> 0%)
ActionDamping d=2: prior suppression of action logits (the served numbers above)
Deterministic: sha256(prompt) seeds sampling; same prompt, same cat

Samples

"can I rub her tummy?" -> " Purrr! "
"did you see that dog" -> " Mraow Mew!"
"How was your day" -> "Mewmew. Mrow!"
"it is 3am" -> " Meow ..."

Usage

Server, tokenizer, eval batteries, and the companion calmer model: https://github.com/tampajohn/meow-lite (MEOW_LITE_ENGINE=v5)

Dataset + full methodology: https://huggingface.co/datasets/tampajohn/meow-lite-v6-dataset

Downloads last month: 238

Safetensors

Model size

6.84M params

Tensor type

F32