Instructions to use cs-552-2026-MandMP/math_model with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use cs-552-2026-MandMP/math_model with Transformers:

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

pipe = pipeline("text-generation", model="cs-552-2026-MandMP/math_model")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("cs-552-2026-MandMP/math_model")
model = AutoModelForCausalLM.from_pretrained("cs-552-2026-MandMP/math_model", 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 cs-552-2026-MandMP/math_model with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "cs-552-2026-MandMP/math_model"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "cs-552-2026-MandMP/math_model",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/cs-552-2026-MandMP/math_model

SGLang

How to use cs-552-2026-MandMP/math_model 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 "cs-552-2026-MandMP/math_model" \
    --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": "cs-552-2026-MandMP/math_model",
		"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 "cs-552-2026-MandMP/math_model" \
        --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": "cs-552-2026-MandMP/math_model",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use cs-552-2026-MandMP/math_model with Docker Model Runner:
```
docker model run hf.co/cs-552-2026-MandMP/math_model
```

math_model

Commit History

Automated MNLP evaluation report (2026-06-03) (#14)

a41c3c3

sohaddad

zechen-nlp commited on Jun 3

Automated MNLP evaluation report (2026-06-01) (#13)

0955d60

sohaddad

zechen-nlp commited on Jun 1

Automated MNLP evaluation report (2026-05-29) (#12)

58b649f

sohaddad

zechen-nlp commited on May 29

concise_sft_v1: SFT on 7899 concise traces from Qwen3-8B (thinking off), 1 epoch on top of dapo_v3

051ffca
verified

sohaddad commited on May 28

Automated MNLP evaluation report (2026-05-28) (#11)

f58c2f0

sohaddad

zechen-nlp commited on May 28

Automated MNLP evaluation report (2026-05-27) (#10)

2c97fe4

sohaddad

zechen-nlp commited on May 27

Automated MNLP evaluation report (2026-05-26) (#9)

e95e6f9

sohaddad

zechen-nlp commited on May 26

Upload dapo_v3_merged_latest weights

46b0423
verified

sohaddad commited on May 25

Add repetition_penalty=1.05

c1721de
verified

sohaddad commited on May 25

Automated MNLP evaluation report (2026-05-25) (#8)

d660f80

sohaddad

zechen-nlp commited on May 25

Automated MNLP evaluation report (2026-05-24) (#7)

068d0e5

sohaddad

zechen-nlp commited on May 24

DAPO RL 1600 steps merged

1418c00
verified

sohaddad commited on May 24

Automated MNLP evaluation report (2026-05-23) (#6)

5a568eb

sohaddad

zechen-nlp commited on May 24

DAPO RL 800 steps merged

5a9ca7a
verified

sohaddad commited on May 22

Upload dapo_v2 merged checkpoint at step 200

88c1571
verified

sohaddad commited on May 21

Automated MNLP evaluation report (2026-05-21) (#5)

335c35a

sohaddad

zechen-nlp commited on May 21

Add checkpoint-580: SFT distill with math system prompt + enable_thinking

6e0bf51
verified

sohaddad commited on May 20

Fix tokenizer_config: set pad_token to <|im_end|>

796af15
verified

sohaddad commited on May 20

Fix chat_template: add enable_thinking=true and math system prompt auto-injection

7b6d984
verified

sohaddad commited on May 20

Automated MNLP evaluation report (2026-05-20) (#4)

ba906aa

sohaddad

zechen-nlp commited on May 20

fix: default enable_thinking to ON by adding else branch for <think> injection

0f88a28
verified

sohaddad commited on May 19

fix: update generation_config.json to match project spec

264aa12
verified

sohaddad commited on May 19

fix: set pad_token_id to 151643 (endoftext)

46c3f56
verified

sohaddad commited on May 19

Fix: restore Qwen3 chat template

95d0da9
verified

sohaddad commited on May 19

Fix: restore Qwen3 chat template

bcc275e
verified

sohaddad commited on May 19

Automated MNLP evaluation report (2026-05-19) (#3)

dc38dbd

sohaddad

zechen-nlp commited on May 19

Fix generation_config: correct pad_token_id, add bos_token_id, remove training-time overrides

3a4ce59
verified

sohaddad commited on May 18

Automated MNLP evaluation report (2026-05-18) (#2)

63c4966

sohaddad

zechen-nlp commited on May 18

Remove training_args.bin (training artifact, not needed for inference)

b455b43
verified

sohaddad commited on May 17

Remove trainer_state.json (training artifact, not needed for inference)

eccf3f4
verified

sohaddad commited on May 17

Remove scheduler.pt (training artifact, not needed for inference)

0b52c5b
verified

sohaddad commited on May 17

Remove rng_state.pth (training artifact, not needed for inference)

9584bd0
verified

sohaddad commited on May 17

Remove optimizer.pt (training artifact, not needed for inference)

0776556
verified

sohaddad commited on May 17

Distilled checkpoint (Qwen3-8B teacher, MATH train, 2 epochs)

6f45883
verified

sohaddad commited on May 17

Automated MNLP evaluation report (2026-05-13) (#1)

b90a696

sohaddad

zechen-nlp commited on May 13

Add thinking_token_budget=1500 to generation config

bec4dc9
verified

sohaddad commited on May 12

Fix system prompt with correct boxed format

650e0ad
verified

sohaddad commited on May 12

Add default system prompt for boxed format compliance

c79a414
verified

sohaddad commited on May 12

Remove training_args.bin artifact

d04b411
verified

sohaddad commited on May 12

Restore max_new_tokens in generation_config

fdeb43f
verified

sohaddad commited on May 12

Remove training artifacts and checkpoint folders

5299acc
verified

sohaddad commited on May 12

Upload folder using huggingface_hub

bf7cb5b
verified

sohaddad commited on May 12

Upload tokenizer

a141847
verified

sohaddad commited on May 12

Upload Qwen3ForCausalLM

609e9b1
verified

sohaddad commited on May 12

initial commit

56002cc
verified

sohaddad commited on May 12

Commit History

Automated MNLP evaluation report (2026-06-03) (#14) a41c3c3

Automated MNLP evaluation report (2026-06-01) (#13) 0955d60

Automated MNLP evaluation report (2026-05-29) (#12) 58b649f

concise_sft_v1: SFT on 7899 concise traces from Qwen3-8B (thinking off), 1 epoch on top of dapo_v3 051ffca verified

Automated MNLP evaluation report (2026-05-28) (#11) f58c2f0

Automated MNLP evaluation report (2026-05-27) (#10) 2c97fe4

Automated MNLP evaluation report (2026-05-26) (#9) e95e6f9

Upload dapo_v3_merged_latest weights 46b0423 verified

Add repetition_penalty=1.05 c1721de verified

Automated MNLP evaluation report (2026-05-25) (#8) d660f80

Automated MNLP evaluation report (2026-05-24) (#7) 068d0e5

DAPO RL 1600 steps merged 1418c00 verified

Automated MNLP evaluation report (2026-05-23) (#6) 5a568eb

DAPO RL 800 steps merged 5a9ca7a verified

Upload dapo_v2 merged checkpoint at step 200 88c1571 verified

Automated MNLP evaluation report (2026-05-21) (#5) 335c35a

Add checkpoint-580: SFT distill with math system prompt + enable_thinking 6e0bf51 verified

Fix tokenizer_config: set pad_token to <|im_end|> 796af15 verified

Fix chat_template: add enable_thinking=true and math system prompt auto-injection 7b6d984 verified

Automated MNLP evaluation report (2026-05-20) (#4) ba906aa

fix: default enable_thinking to ON by adding else branch for <think> injection 0f88a28 verified

fix: update generation_config.json to match project spec 264aa12 verified

fix: set pad_token_id to 151643 (endoftext) 46c3f56 verified

Fix: restore Qwen3 chat template 95d0da9 verified

Fix: restore Qwen3 chat template bcc275e verified

Automated MNLP evaluation report (2026-05-19) (#3) dc38dbd

Fix generation_config: correct pad_token_id, add bos_token_id, remove training-time overrides 3a4ce59 verified

Automated MNLP evaluation report (2026-05-18) (#2) 63c4966

Remove training_args.bin (training artifact, not needed for inference) b455b43 verified

Remove trainer_state.json (training artifact, not needed for inference) eccf3f4 verified

Remove scheduler.pt (training artifact, not needed for inference) 0b52c5b verified

Remove rng_state.pth (training artifact, not needed for inference) 9584bd0 verified

Remove optimizer.pt (training artifact, not needed for inference) 0776556 verified

Distilled checkpoint (Qwen3-8B teacher, MATH train, 2 epochs) 6f45883 verified

Automated MNLP evaluation report (2026-05-13) (#1) b90a696

Add thinking_token_budget=1500 to generation config bec4dc9 verified

Fix system prompt with correct boxed format 650e0ad verified

Add default system prompt for boxed format compliance c79a414 verified

Remove training_args.bin artifact d04b411 verified

Restore max_new_tokens in generation_config fdeb43f verified

Remove training artifacts and checkpoint folders 5299acc verified

Upload folder using huggingface_hub bf7cb5b verified

Upload tokenizer a141847 verified

Upload Qwen3ForCausalLM 609e9b1 verified

initial commit 56002cc verified

Automated MNLP evaluation report (2026-06-03) (#14)

a41c3c3

Automated MNLP evaluation report (2026-06-01) (#13)

0955d60

Automated MNLP evaluation report (2026-05-29) (#12)

58b649f

concise_sft_v1: SFT on 7899 concise traces from Qwen3-8B (thinking off), 1 epoch on top of dapo_v3

051ffca
verified

Automated MNLP evaluation report (2026-05-28) (#11)

f58c2f0

Automated MNLP evaluation report (2026-05-27) (#10)

2c97fe4

Automated MNLP evaluation report (2026-05-26) (#9)

e95e6f9

Upload dapo_v3_merged_latest weights

46b0423
verified

Add repetition_penalty=1.05

c1721de
verified

Automated MNLP evaluation report (2026-05-25) (#8)

d660f80

Automated MNLP evaluation report (2026-05-24) (#7)

068d0e5

DAPO RL 1600 steps merged

1418c00
verified

Automated MNLP evaluation report (2026-05-23) (#6)

5a568eb

DAPO RL 800 steps merged

5a9ca7a
verified

Upload dapo_v2 merged checkpoint at step 200

88c1571
verified

Automated MNLP evaluation report (2026-05-21) (#5)

335c35a

Add checkpoint-580: SFT distill with math system prompt + enable_thinking

6e0bf51
verified

Fix tokenizer_config: set pad_token to <|im_end|>

796af15
verified

Fix chat_template: add enable_thinking=true and math system prompt auto-injection

7b6d984
verified

Automated MNLP evaluation report (2026-05-20) (#4)

ba906aa

fix: default enable_thinking to ON by adding else branch for <think> injection

0f88a28
verified

fix: update generation_config.json to match project spec

264aa12
verified

fix: set pad_token_id to 151643 (endoftext)

46c3f56
verified

Fix: restore Qwen3 chat template

95d0da9
verified

Fix: restore Qwen3 chat template

bcc275e
verified

Automated MNLP evaluation report (2026-05-19) (#3)

dc38dbd

Fix generation_config: correct pad_token_id, add bos_token_id, remove training-time overrides

3a4ce59
verified

Automated MNLP evaluation report (2026-05-18) (#2)

63c4966

Remove training_args.bin (training artifact, not needed for inference)

b455b43
verified

Remove trainer_state.json (training artifact, not needed for inference)

eccf3f4
verified

Remove scheduler.pt (training artifact, not needed for inference)

0b52c5b
verified

Remove rng_state.pth (training artifact, not needed for inference)

9584bd0
verified

Remove optimizer.pt (training artifact, not needed for inference)

0776556
verified

Distilled checkpoint (Qwen3-8B teacher, MATH train, 2 epochs)

6f45883
verified

Automated MNLP evaluation report (2026-05-13) (#1)

b90a696

Add thinking_token_budget=1500 to generation config

bec4dc9
verified

Fix system prompt with correct boxed format

650e0ad
verified

Add default system prompt for boxed format compliance

c79a414
verified

Remove training_args.bin artifact

d04b411
verified

Restore max_new_tokens in generation_config

fdeb43f
verified

Remove training artifacts and checkpoint folders

5299acc
verified

Upload folder using huggingface_hub

bf7cb5b
verified

Upload tokenizer

a141847
verified

Upload Qwen3ForCausalLM

609e9b1
verified

initial commit

56002cc
verified