Instructions to use sfanm/d24-midtrain-v1base-mathheavy-3.7B with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use sfanm/d24-midtrain-v1base-mathheavy-3.7B with Transformers:

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

pipe = pipeline("text-generation", model="sfanm/d24-midtrain-v1base-mathheavy-3.7B")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("sfanm/d24-midtrain-v1base-mathheavy-3.7B")
model = AutoModelForCausalLM.from_pretrained("sfanm/d24-midtrain-v1base-mathheavy-3.7B", device_map="auto")
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 Settings

vLLM

How to use sfanm/d24-midtrain-v1base-mathheavy-3.7B with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "sfanm/d24-midtrain-v1base-mathheavy-3.7B"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "sfanm/d24-midtrain-v1base-mathheavy-3.7B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/sfanm/d24-midtrain-v1base-mathheavy-3.7B

SGLang

How to use sfanm/d24-midtrain-v1base-mathheavy-3.7B 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 "sfanm/d24-midtrain-v1base-mathheavy-3.7B" \
    --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": "sfanm/d24-midtrain-v1base-mathheavy-3.7B",
		"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 "sfanm/d24-midtrain-v1base-mathheavy-3.7B" \
        --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": "sfanm/d24-midtrain-v1base-mathheavy-3.7B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use sfanm/d24-midtrain-v1base-mathheavy-3.7B with Docker Model Runner:
```
docker model run hf.co/sfanm/d24-midtrain-v1base-mathheavy-3.7B
```

d24-midtrain-v1base-mathheavy-3.7B / README.md

sfanm

Upload README.md with huggingface_hub

a6f0b97 verified 4 months ago

preview code

Raw History Blame Contribute Delete

1.66 kB

	---
	license: other
	language: en
	library_name: transformers
	pipeline_tag: text-generation
	tags:
	- nanochat
	- nemotron
	- from-scratch
	- perlmutter
	- gpt2-tokenizer
	---
	# d24-midtrain-v1base-mathheavy-3.7B

	v1-base math-heavy-midtrained BASE LM (pre-SFT).

	nanochat-style depth-24 decoder — 24 layers × 1536 hidden × 12 heads, SwiGLU / RoPE / RMSNorm, tied embeddings, GPT-2 BPE vocab (50304), 0.757B params, 2048-token context.

	Lineage. v1 pretrain (5.84B ClimbMix) → math-heavy midtrain 3.7B (FineMath/OpenMath/MetaMath/OpenThoughts + ClimbMix anchor).

	Metrics. Base checkpoint (pre-SFT) — evaluate after SFT. Corresponding SFT: `d24-sft-v1base-mathheavy-3.7B` (GSM8K 5.46%).

	## Use (base LM)

	This is a base language model (post-midtrain, pre-SFT) — use it for text continuation, not chat. EOS is the GPT-2 `<\|endoftext\|>` (`50256`). For a chat model, use the `d24-sft-*` checkpoints.

	```python
	from transformers import AutoModelForCausalLM, AutoTokenizer
	mid = "sfanm/d24-midtrain-v1base-mathheavy-3.7B"
	tok = AutoTokenizer.from_pretrained(mid)
	model = AutoModelForCausalLM.from_pretrained(mid, torch_dtype="bfloat16", device_map="auto")
	inputs = tok("The derivative of x**2 is", return_tensors="pt").to(model.device)
	print(tok.decode(model.generate(**inputs, max_new_tokens=128)[0], skip_special_tokens=True))
	```

	Research checkpoint from a from-scratch nanochat-d24 replication (pretrain → midtrain → SFT → RL) on NERSC Perlmutter. Trained on third-party corpora (ClimbMix, FineMath, OpenMath, MetaMath, OpenThoughts, OLMo-3 Dolmino, SmolTalk, …) — see those datasets' licenses; provided as-is for research.