Instructions to use vtava/TinyCeNN-LM-Sharded-MoE-Top2 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use vtava/TinyCeNN-LM-Sharded-MoE-Top2 with Transformers:

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

pipe = pipeline("text-generation", model="vtava/TinyCeNN-LM-Sharded-MoE-Top2")

# Load model directly
from transformers import AutoModel
model = AutoModel.from_pretrained("vtava/TinyCeNN-LM-Sharded-MoE-Top2", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use vtava/TinyCeNN-LM-Sharded-MoE-Top2 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "vtava/TinyCeNN-LM-Sharded-MoE-Top2"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "vtava/TinyCeNN-LM-Sharded-MoE-Top2",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/vtava/TinyCeNN-LM-Sharded-MoE-Top2

SGLang

How to use vtava/TinyCeNN-LM-Sharded-MoE-Top2 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 "vtava/TinyCeNN-LM-Sharded-MoE-Top2" \
    --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": "vtava/TinyCeNN-LM-Sharded-MoE-Top2",
		"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 "vtava/TinyCeNN-LM-Sharded-MoE-Top2" \
        --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": "vtava/TinyCeNN-LM-Sharded-MoE-Top2",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use vtava/TinyCeNN-LM-Sharded-MoE-Top2 with Docker Model Runner:
```
docker model run hf.co/vtava/TinyCeNN-LM-Sharded-MoE-Top2
```

TinyCeNN-LM Sharded MoE Top-2

Transformer-free CeNN student using a parameter-neutral routed FFN.

8 shards × 96 inner channels
Top-2 routing
trainable parameters: 481,729
overhead vs plain CeNN: 0.320%
held-out tokens: 65,536
best CE: 5.115179
teacher CE: 4.234912
teacher-gap recovery: 84.55%

Downloads last month: -; Downloads are not tracked for this model. How to track

Model tree for vtava/TinyCeNN-LM-Sharded-MoE-Top2

Base model

arnir0/Tiny-LLM

Finetuned

(10)

this model