Instructions to use cs-552-2026-kth/math_model with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use cs-552-2026-kth/math_model with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("text-generation", model="cs-552-2026-kth/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-kth/math_model") model = AutoModelForCausalLM.from_pretrained("cs-552-2026-kth/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-kth/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-kth/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-kth/math_model", "messages": [ { "role": "user", "content": "What is the capital of France?" } ] }'Use Docker
docker model run hf.co/cs-552-2026-kth/math_model
- SGLang
How to use cs-552-2026-kth/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-kth/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-kth/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-kth/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-kth/math_model", "messages": [ { "role": "user", "content": "What is the capital of France?" } ] }' - Docker Model Runner
How to use cs-552-2026-kth/math_model with Docker Model Runner:
docker model run hf.co/cs-552-2026-kth/math_model
Update Automated MNLP evaluation report (2026-05-17)
Browse files- EVAL_REPORT.md +70 -8
EVAL_REPORT.md
CHANGED
|
@@ -2,7 +2,7 @@
|
|
| 2 |
|
| 3 |
- **Model repo:** [`cs-552-2026-kth/math_model`](https://huggingface.co/cs-552-2026-kth/math_model)
|
| 4 |
- **Owner(s):** group **kth**
|
| 5 |
-
- **Generated at:** 2026-05-
|
| 6 |
- **Pipeline:** [mnlp-project-ci](https://github.com/eric11eca/mnlp-project-ci)
|
| 7 |
|
| 8 |
_This PR is opened automatically by the course CI. It is **non-blocking** — you do not need to merge it. The next nightly run will refresh this file._
|
|
@@ -11,17 +11,79 @@ _This PR is opened automatically by the course CI. It is **non-blocking** — yo
|
|
| 11 |
|
| 12 |
| Benchmark | Accuracy | Status |
|
| 13 |
|---|---:|---|
|
| 14 |
-
| Math |
|
| 15 |
| Knowledge | — | not run |
|
| 16 |
| Multilingual | — | not run |
|
| 17 |
| Safety | — | not run |
|
| 18 |
|
| 19 |
-
##
|
| 20 |
|
| 21 |
-
|
| 22 |
|
| 23 |
-
|
| 24 |
-
generation_config.json not found: 404 Client Error. (Request ID: Root=1-6a07f219-31df42fa24af8d9830de3e10;1281fcc3-f3c3-4f28-8c3d-eaa5f9a55145)
|
| 25 |
|
| 26 |
-
|
| 27 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 2 |
|
| 3 |
- **Model repo:** [`cs-552-2026-kth/math_model`](https://huggingface.co/cs-552-2026-kth/math_model)
|
| 4 |
- **Owner(s):** group **kth**
|
| 5 |
+
- **Generated at:** 2026-05-17T04:52:19+00:00 (UTC)
|
| 6 |
- **Pipeline:** [mnlp-project-ci](https://github.com/eric11eca/mnlp-project-ci)
|
| 7 |
|
| 8 |
_This PR is opened automatically by the course CI. It is **non-blocking** — you do not need to merge it. The next nightly run will refresh this file._
|
|
|
|
| 11 |
|
| 12 |
| Benchmark | Accuracy | Status |
|
| 13 |
|---|---:|---|
|
| 14 |
+
| Math | 0.3100 | ok |
|
| 15 |
| Knowledge | — | not run |
|
| 16 |
| Multilingual | — | not run |
|
| 17 |
| Safety | — | not run |
|
| 18 |
|
| 19 |
+
## Sample completions
|
| 20 |
|
| 21 |
+
_Prompts are intentionally omitted to avoid revealing benchmark contents. For multi-completion problems, only one completion is shown per sample._
|
| 22 |
|
| 23 |
+
### Math
|
|
|
|
| 24 |
|
| 25 |
+
**Correct** (1 shown)
|
| 26 |
+
|
| 27 |
+
- **reference**: `4`
|
| 28 |
+
- **overall** (8/8 completions correct)
|
| 29 |
+
- **extracted** (✓): `4`
|
| 30 |
+
- **completion**:
|
| 31 |
+
|
| 32 |
+
```text
|
| 33 |
+
<think>
|
| 34 |
+
We can rewrite the equation as:
|
| 35 |
+
\[ 10^x = 10 + 9990 \Rightarrow 10^x = 10000 \Rightarrow 10^x = 10^4 \]
|
| 36 |
+
|
| 37 |
+
Since the bases are the same, we can equate the exponents:
|
| 38 |
+
\[ x = 4 \]
|
| 39 |
+
|
| 40 |
+
The answer is:
|
| 41 |
+
\[
|
| 42 |
+
</think>
|
| 43 |
+
|
| 44 |
+
Therefore, the answer is \boxed{4}.
|
| 45 |
+
```
|
| 46 |
+
|
| 47 |
+
**Incorrect** (1 shown)
|
| 48 |
+
|
| 49 |
+
- **reference**: `2 - \frac{\pi}{2}`
|
| 50 |
+
- **overall** (0/8 completions correct)
|
| 51 |
+
- **extracted** (✗): `\frac{1}{2}`
|
| 52 |
+
- **completion**:
|
| 53 |
+
|
| 54 |
+
```text
|
| 55 |
+
<think>
|
| 56 |
+
Let $P$ be the probability that Derek eventually becomes debt-free. We can derive a recursive equation for $P$.
|
| 57 |
+
|
| 58 |
+
When Derek tosses the coin, there are two possible outcomes:
|
| 59 |
+
|
| 60 |
+
1. Heads: He reduces his debt by 1 MIT. This means his debt becomes $\pi - 1$ MITs.
|
| 61 |
+
2. Tails: He reduces his debt by 2 MITs. This means his debt becomes $\pi - 2$ MITs.
|
| 62 |
+
|
| 63 |
+
However, if his debt becomes nonpositive, he is
|
| 64 |
+
|
| 65 |
+
... [truncated 540 chars] ...
|
| 66 |
+
|
| 67 |
+
ly becomes debt-free given a debt of $\pi - 1$ MITs, and $P_{\pi - 2}$ is the probability that Derek eventually becomes debt-free given a debt of $\pi - 2$ MITs.
|
| 68 |
+
|
| 69 |
+
We can continue this process recursively, but we notice a pattern. The recursive equation can be rewritten as:
|
| 70 |
+
|
| 71 |
+
\[ P = \frac{1}{2} \left( P_{\pi - 1} + P_{\pi - 2} \right) \]
|
| 72 |
+
|
| 73 |
+
Using the same logic, we can derive the following recursive equation:
|
| 74 |
+
|
| 75 |
+
\[ P = \frac{1}{2} \left( \frac{1}{2} \left( P_{\pi - 2} + P_{\pi - 3} \right) + \frac{1}{2} \left( P_{\pi - 3} + P_{\pi - 4} \right) \right) \]
|
| 76 |
+
|
| 77 |
+
And so on.
|
| 78 |
+
|
| 79 |
+
Notice that the recursive equation can be simplified to:
|
| 80 |
+
|
| 81 |
+
\[ P = \frac{1}{2} \left( \frac{1}{2} \left( \frac{1}{2} \left( \cdots \right) + \frac{1}{2} \left( \frac{1}{2} \left( \cdots \right) \right) \right) + \frac{1}{2} \left( \frac{1}{2} \left( \cdots \right) + \frac{1}{2} \left( \frac{1}{2} \left( \cdots \right) \right) \right) \right) + \cdots \right) \]
|
| 82 |
+
|
| 83 |
+
This is a geometric series with first term $\frac{1}{2}$ and common ratio $\frac{1}{2}$. The sum of the series is:
|
| 84 |
+
|
| 85 |
+
\[ P = \frac{\frac{1}{2}}{1 - \frac{1}{2}} = \frac{1}{2} \div \frac{1}{2} = \frac{1}{2} \cdot 2 =
|
| 86 |
+
</think>
|
| 87 |
+
|
| 88 |
+
Therefore, the answer is \boxed{\frac{1}{2}}.
|
| 89 |
+
```
|