Instructions to use rahulshetty/hybrid-tiny-wikitext-poc with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use rahulshetty/hybrid-tiny-wikitext-poc with Transformers:

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

pipe = pipeline("text-generation", model="rahulshetty/hybrid-tiny-wikitext-poc")

# Load model directly
from transformers import AutoModelForCausalLM
model = AutoModelForCausalLM.from_pretrained("rahulshetty/hybrid-tiny-wikitext-poc", dtype="auto")

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use rahulshetty/hybrid-tiny-wikitext-poc with vLLM:

Install from pip and serve model

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

Use Docker

docker model run hf.co/rahulshetty/hybrid-tiny-wikitext-poc

SGLang

How to use rahulshetty/hybrid-tiny-wikitext-poc 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 "rahulshetty/hybrid-tiny-wikitext-poc" \
    --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": "rahulshetty/hybrid-tiny-wikitext-poc",
		"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 "rahulshetty/hybrid-tiny-wikitext-poc" \
        --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": "rahulshetty/hybrid-tiny-wikitext-poc",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use rahulshetty/hybrid-tiny-wikitext-poc with Docker Model Runner:
```
docker model run hf.co/rahulshetty/hybrid-tiny-wikitext-poc
```

Hybrid Tiny LM POC on WikiText-2

Proof-of-concept causal language model combining multiple token-mixing approaches:

local/full causal self-attention for exact token lookup
gated causal depthwise convolution for local pattern mixing
input-gated diagonal recurrent mixer as a tiny SSM/RWKV-like compressed memory
hybrid blocks that run attention + conv + recurrence in parallel and fuse them

Dataset: Salesforce/wikitext, config wikitext-2-raw-v1. Tokenizer: openai-community/gpt2.

POC metrics

{
  "eval_loss": 7.388123512268066,
  "eval_runtime": 9.5676,
  "eval_samples_per_second": 26.757,
  "eval_steps_per_second": 3.345,
  "epoch": 0.78125,
  "perplexity": 1616.6696041601817,
  "train_loss": 7.8487934923172,
  "params": 10403040
}

This is intentionally tiny and trained briefly as an architecture POC, not a competitive LM.

Reproduce

pip install "transformers>=4.54.0" datasets torch accelerate trackio
python hybrid_lm_poc.py \
  --max_steps 200 \
  --max_train_samples 2048 \
  --max_eval_samples 256 \
  --batch_size 8 \
  --d_model 96 \
  --n_layer 6 \
  --block_size 64 \
  --learning_rate 8e-4

Sample generation after the short run is in sample_generation.txt.

Generated by ML Intern

This model repository was generated by ML Intern, an agent for machine learning research and development on the Hugging Face Hub.

Try ML Intern: https://smolagents-ml-intern.hf.space
Source code: https://github.com/huggingface/ml-intern

Usage

from transformers import AutoModelForCausalLM, AutoTokenizer

model_id = 'rahulshetty/hybrid-tiny-wikitext-poc'
tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForCausalLM.from_pretrained(model_id)

For non-causal architectures, replace AutoModelForCausalLM with the appropriate AutoModel class.

Downloads last month: -

Safetensors

Model size

10.4M params

Tensor type

F32

rahulshetty
/

hybrid-tiny-wikitext-poc

Hybrid Tiny LM POC on WikiText-2

POC metrics

Reproduce

Generated by ML Intern

Usage

Dataset used to train rahulshetty/hybrid-tiny-wikitext-poc