Instructions to use sfanm/d24-sft-v2-reasoning-3.7B with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use sfanm/d24-sft-v2-reasoning-3.7B with Transformers:

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

pipe = pipeline("text-generation", model="sfanm/d24-sft-v2-reasoning-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-sft-v2-reasoning-3.7B")
model = AutoModelForCausalLM.from_pretrained("sfanm/d24-sft-v2-reasoning-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-sft-v2-reasoning-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-sft-v2-reasoning-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-sft-v2-reasoning-3.7B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/sfanm/d24-sft-v2-reasoning-3.7B

SGLang

How to use sfanm/d24-sft-v2-reasoning-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-sft-v2-reasoning-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-sft-v2-reasoning-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-sft-v2-reasoning-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-sft-v2-reasoning-3.7B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use sfanm/d24-sft-v2-reasoning-3.7B with Docker Model Runner:
```
docker model run hf.co/sfanm/d24-sft-v2-reasoning-3.7B
```

sfanm commited on Jun 12

Commit

295ff4f

verified ·

1 Parent(s): ac09918

Upload README.md with huggingface_hub

Browse files

Files changed (1) hide show

README.md +14 -4

README.md CHANGED Viewed

@@ -20,12 +20,22 @@ nanochat-style **depth-24** decoder — 24 layers × 1536 hidden × 12 heads, Sw
 **Metrics.** GSM8K (greedy, full 1319): **9.86%** · SFT val lm-loss 0.162.
-## Load
 ```python
 from transformers import AutoModelForCausalLM, AutoTokenizer
-m = "sfanm/d24-sft-v2-reasoning-3.7B"
-tok = AutoTokenizer.from_pretrained(m)
-model = AutoModelForCausalLM.from_pretrained(m, torch_dtype="bfloat16")
 ```
 *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.*

 **Metrics.** GSM8K (greedy, full 1319): **9.86%** · SFT val lm-loss 0.162.
+## Use (chat)
+This is a **chat model** (ChatML). The turn terminator it emits is the literal string **`<|im_end|>`** — which is **not** the `eos_token_id` (`50256` = `<|endoftext|>`) and is not even a single token. You **must** stop on the `<|im_end|>` string or generation will not stop:
 ```python
 from transformers import AutoModelForCausalLM, AutoTokenizer
+mid = "sfanm/d24-sft-v2-reasoning-3.7B"
+tok = AutoTokenizer.from_pretrained(mid)
+model = AutoModelForCausalLM.from_pretrained(mid, torch_dtype="bfloat16", device_map="auto")
+msgs = [{"role": "user", "content": "Natalia sold clips to 48 friends in April and half as many in May. How many total?"}]
+ids = tok.apply_chat_template(msgs, add_generation_prompt=True, return_tensors="pt").to(model.device)
+out = model.generate(ids, max_new_tokens=512, do_sample=False, stop_strings=["<|im_end|>"], tokenizer=tok)
+print(tok.decode(out[0, ids.shape[1]:], skip_special_tokens=False))
 ```
+> Without `stop_strings=["<|im_end|>"]` the model rambles to `max_new_tokens`: the configured `eos_token_id` (50256) is the GPT-2 *document* EOS, which a chat turn does not end with. For vLLM, pass `stop=["<|im_end|>"]`.
 *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.*