Instructions to use HuggingFaceTB/nanowhale-100m-base with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use HuggingFaceTB/nanowhale-100m-base with Transformers:

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

pipe = pipeline("text-generation", model="HuggingFaceTB/nanowhale-100m-base", trust_remote_code=True)
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("HuggingFaceTB/nanowhale-100m-base", trust_remote_code=True)
model = AutoModelForCausalLM.from_pretrained("HuggingFaceTB/nanowhale-100m-base", trust_remote_code=True)
messages = [
    {"role": "user", "content": "Who are you?"},
]
inputs = tokenizer.apply_chat_template(
	messages,
	add_generation_prompt=True,
	tokenize=True,
	return_dict=True,
	return_tensors="pt",
).to(model.device)

outputs = model.generate(**inputs, max_new_tokens=40)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use HuggingFaceTB/nanowhale-100m-base with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "HuggingFaceTB/nanowhale-100m-base"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "HuggingFaceTB/nanowhale-100m-base",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/HuggingFaceTB/nanowhale-100m-base

SGLang

How to use HuggingFaceTB/nanowhale-100m-base 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 "HuggingFaceTB/nanowhale-100m-base" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "HuggingFaceTB/nanowhale-100m-base",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

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 "HuggingFaceTB/nanowhale-100m-base" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "HuggingFaceTB/nanowhale-100m-base",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use HuggingFaceTB/nanowhale-100m-base with Docker Model Runner:
```
docker model run hf.co/HuggingFaceTB/nanowhale-100m-base
```

nanowhale-100m-base / README.md

cmpatino HF Staff

Add YAML metadata to model card

8ea74cc verified 2 days ago

preview code

raw

history blame contribute delete

4 kB

	---
	license: apache-2.0
	language:
	- en
	library_name: transformers
	tags:
	- deepseek
	- moe
	- causal-lm
	- pretrained
	datasets:
	- HuggingFaceFW/fineweb-edu
	pipeline_tag: text-generation
	model-index:
	- name: nanowhale-100m-base
	results: []
	---

	# nanowhale-100m-base 🐳

	A small ~110M parameter language model implementing the DeepSeek-V4 architecture from scratch. This is the pretrained base model — see [HuggingFaceTB/nanowhale-100m](https://huggingface.co/HuggingFaceTB/nanowhale-100m) for the SFT/chat version.

	Training code: [github.com/huggingface/nanowhale](https://github.com/huggingface/nanowhale)

	## Architecture

	This model implements key DeepSeek-V4 innovations at a miniature scale:

	\| Component \| Details \|
	\|---\|---\|
	\| Parameters \| ~110M total (41M embeddings, 69M non-embedding) \|
	\| Hidden size \| 320 \|
	\| Layers \| 8 \|
	\| Attention heads \| 8 (1 KV head — MQA-style) \|
	\| Head dim \| 96 (32 RoPE + 64 NoPE) \|
	\| MLA \| q_lora_rank=160, o_groups=2, o_lora_rank=80 \|
	\| MoE \| 4 routed experts + 1 shared, top-2 routing \|
	\| Expert FFN \| SwiGLU, intermediate_size=640 \|
	\| Routing \| sqrtsoftplus scoring, noaux_tc method \|
	\| Hyper-Connections \| hc_mult=4, Sinkhorn routing (2 iters) \|
	\| MTP \| 1 next-token prediction layer \|
	\| Vocab \| 129,280 (DeepSeek-V4 tokenizer) \|
	\| Context \| 2,048 tokens \|

	### DeepSeek-V4 Features Implemented

	- Multi-head Latent Attention (MLA): Compressed KV cache via latent projections
	- Mixture of Experts (MoE): Sparse activation — only 2 of 4 experts per token
	- Hyper-Connections: Multi-copy hidden states with learned Sinkhorn routing replacing residual connections
	- SwiGLU FFN with configurable limit
	- Grouped output projection (o_groups)

	## Training

	- Dataset: [HuggingFaceFW/fineweb-edu](https://huggingface.co/datasets/HuggingFaceFW/fineweb-edu) (streaming)
	- Steps: 5,000
	- Tokens seen: ~2.6B
	- Batch size: 8 × 4 gradient accumulation = 32 effective
	- Sequence length: 2,048
	- Learning rate: 6e-4, cosine schedule, 3% warmup
	- Optimizer: AdamW (β1=0.9, β2=0.95, weight_decay=0.1)
	- Precision: bf16 mixed precision
	- Hardware: 1× NVIDIA H100 80GB

	### Training Metrics

	\| Metric \| Value \|
	\|---\|---\|
	\| Final loss \| ~5.3 (cross-entropy) \|
	\| Final entropy \| 3.77 \|
	\| Token accuracy \| 33.8% \|

	## Usage

	```python
	import torch
	from safetensors.torch import load_file
	from transformers import AutoConfig, AutoModelForCausalLM, AutoTokenizer
	from huggingface_hub import hf_hub_download

	# Load model (recommended: manual load for reliability)
	config = AutoConfig.from_pretrained("HuggingFaceTB/nanowhale-100m-base", trust_remote_code=True)
	model = AutoModelForCausalLM.from_config(config, trust_remote_code=True).float()

	# Download and load weights
	weights_path = hf_hub_download("HuggingFaceTB/nanowhale-100m-base", "model.safetensors")
	state_dict = load_file(weights_path)
	model.load_state_dict(state_dict, strict=True)
	model = model.cuda().eval()

	tokenizer = AutoTokenizer.from_pretrained("HuggingFaceTB/nanowhale-100m-base")

	# Generate
	input_ids = tokenizer.encode("The meaning of life is", return_tensors="pt").cuda()
	output = model.generate(input_ids, max_new_tokens=100, temperature=0.7, top_p=0.9,
	pad_token_id=tokenizer.eos_token_id)
	print(tokenizer.decode(output[0], skip_special_tokens=True))
	```

	## Limitations

	- Small model: 110M params with 129K vocab means ~37% of parameters are in embeddings, limiting model capacity
	- Limited training: Only 5K steps / 2.6B tokens — significantly undertrained compared to production models
	- Pretrained only: This is a base model without instruction tuning. Outputs are language-model completions, not conversations.
	- bf16 NaN: Use fp32 — the Hyper-Connections architecture produces values that overflow bf16 range at this scale.
	- Custom architecture: Requires `trust_remote_code=True`

	## License

	Apache-2.0