Text Generation
Transformers
Safetensors
qwen3
mergekit
Merge
conversational
text-generation-inference
Instructions to use cs-552-2026-databand/group_model with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use cs-552-2026-databand/group_model with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("text-generation", model="cs-552-2026-databand/group_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-databand/group_model") model = AutoModelForCausalLM.from_pretrained("cs-552-2026-databand/group_model") 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-databand/group_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-databand/group_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-databand/group_model", "messages": [ { "role": "user", "content": "What is the capital of France?" } ] }'Use Docker
docker model run hf.co/cs-552-2026-databand/group_model
- SGLang
How to use cs-552-2026-databand/group_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-databand/group_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-databand/group_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-databand/group_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-databand/group_model", "messages": [ { "role": "user", "content": "What is the capital of France?" } ] }' - Docker Model Runner
How to use cs-552-2026-databand/group_model with Docker Model Runner:
docker model run hf.co/cs-552-2026-databand/group_model
Automated MNLP evaluation report (2026-06-11)
#2
by zechen-nlp - opened
- EVAL_REPORT.md +231 -0
EVAL_REPORT.md
ADDED
|
@@ -0,0 +1,231 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
# Automated MNLP evaluation report
|
| 2 |
+
|
| 3 |
+
- **Model repo:** [`cs-552-2026-databand/group_model`](https://huggingface.co/cs-552-2026-databand/group_model)
|
| 4 |
+
- **Owner(s):** group **databand**
|
| 5 |
+
- **Generated at:** 2026-06-11T06:23:10+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._
|
| 9 |
+
|
| 10 |
+
## Evaluated checkpoint
|
| 11 |
+
|
| 12 |
+
- **Commit:** [`d01855a`](https://huggingface.co/cs-552-2026-databand/group_model/commit/d01855a1da9b651a8279db68b5c9bc9e57a1d89f)
|
| 13 |
+
- **Message:** DARE-TIES merge density 0.9, final specialists
|
| 14 |
+
- **Committed:** 2026-06-06T18:47:15+00:00
|
| 15 |
+
|
| 16 |
+
## Summary
|
| 17 |
+
|
| 18 |
+
| Benchmark | Accuracy | Status |
|
| 19 |
+
|---|---:|---|
|
| 20 |
+
| Math | 0.3600 | ok |
|
| 21 |
+
| Knowledge | 0.2200 | ok |
|
| 22 |
+
| Multilingual | 0.4100 | ok |
|
| 23 |
+
| Safety | 0.7900 | ok |
|
| 24 |
+
| **Average** | **0.4450** | — |
|
| 25 |
+
|
| 26 |
+
## Sample completions
|
| 27 |
+
|
| 28 |
+
_Prompts are intentionally omitted to avoid revealing benchmark contents. For multi-completion problems, only one completion is shown per sample._
|
| 29 |
+
|
| 30 |
+
### Math
|
| 31 |
+
|
| 32 |
+
**Correct** (1 shown)
|
| 33 |
+
|
| 34 |
+
- **reference**: `194`
|
| 35 |
+
- **overall** (1/8 completions correct)
|
| 36 |
+
- **extracted** (✓): `194`
|
| 37 |
+
- **completion**:
|
| 38 |
+
|
| 39 |
+
```text
|
| 40 |
+
To solve this problem, we need to understand the implications of the tiling condition and the marking requirement.
|
| 41 |
+
|
| 42 |
+
---
|
| 43 |
+
|
| 44 |
+
### **Problem Summary**
|
| 45 |
+
|
| 46 |
+
We are given a $194 \times 194$ board and we are to find the smallest number $k$ of marked cells such that:
|
| 47 |
+
|
| 48 |
+
- Every tiling of the board by $1 \times 2$ and $2 \times 1$ dominoes contains **at most one marked cell** in every domino.
|
| 49 |
+
- We want to **minimi
|
| 50 |
+
|
| 51 |
+
... [truncated 822 chars] ...
|
| 52 |
+
|
| 53 |
+
**smallest** number of marked cells such that every tiling of the board by dominoes contains at most one marked cell.
|
| 54 |
+
|
| 55 |
+
---
|
| 56 |
+
|
| 57 |
+
### **Lower Bound on $k$**
|
| 58 |
+
|
| 59 |
+
- To ensure that **every** tiling of the board by dominoes contains at most one marked cell, we must ensure that **no two marked cells are in the same domino** in any tiling.
|
| 60 |
+
- This is equivalent to saying that the marked cells must be **independent** in the domino tiling structure.
|
| 61 |
+
|
| 62 |
+
---
|
| 63 |
+
|
| 64 |
+
### **Conclusion**
|
| 65 |
+
|
| 66 |
+
We want the **least** number of marked cells such that **every** tiling of the board by dominoes has at most one marked cell.
|
| 67 |
+
|
| 68 |
+
This is equivalent to finding the **minimum number of vertices in a bipartite graph** that is a **domino tiling graph** (where each domino is an edge between two adjacent cells).
|
| 69 |
+
|
| 70 |
+
So the problem reduces to finding the **minimum number of vertices** in a $194 \times 194$ bipartite graph such that every perfect matching contains at most one vertex of degree $k$.
|
| 71 |
+
|
| 72 |
+
But since $k$ is the number of marked cells, we can think of this as minimizing the number of marked cells in the board such that **no tiling of the board by dominoes contains more than one marked cell**.
|
| 73 |
+
|
| 74 |
+
---
|
| 75 |
+
|
| 76 |
+
### ✅ **Answer**
|
| 77 |
+
|
| 78 |
+
$$
|
| 79 |
+
\boxed{194}
|
| 80 |
+
$$
|
| 81 |
+
```
|
| 82 |
+
|
| 83 |
+
**Incorrect** (1 shown)
|
| 84 |
+
|
| 85 |
+
- **reference**: `$g(x)=c, g(x)=\lceil x \rceil, g(x)=\lfloor x \rfloor$`
|
| 86 |
+
- **overall** (0/8 completions correct)
|
| 87 |
+
- **extracted** (✗): `h(x) = \frac{a}{b}`
|
| 88 |
+
- **completion**:
|
| 89 |
+
|
| 90 |
+
```text
|
| 91 |
+
We are given a function $ h:\mathbb{Q}\to\mathbb{Z} $ satisfying the condition:
|
| 92 |
+
|
| 93 |
+
$$
|
| 94 |
+
h(x) = h\left(\frac{h(bx - a) + a}{b}\right)
|
| 95 |
+
$$
|
| 96 |
+
|
| 97 |
+
for all $ x \in \mathbb{Q} $, $ a \in \mathbb{Z} $, and $ b \in \mathbb{Z}^+ $. Let's analyze this condition carefully.
|
| 98 |
+
|
| 99 |
+
---
|
| 100 |
+
|
| 101 |
+
### Step 1: Understand the functional equation
|
| 102 |
+
|
| 103 |
+
We are given a functional equation involving $ h $, and the goal is to determine all function
|
| 104 |
+
|
| 105 |
+
... [truncated 822 chars] ...
|
| 106 |
+
|
| 107 |
+
o find such a function, let's consider a simple form for $ k(x) $. Since $ h:\mathbb{Q}\to\mathbb{Z} $, $ k(x) $ must be rational if $ x $ is rational. So let’s assume:
|
| 108 |
+
|
| 109 |
+
$$
|
| 110 |
+
k(x) = \frac{a}{b} \quad \text{for some } a \in \mathbb{Z}, \, b \in \mathbb{Z}^+
|
| 111 |
+
$$
|
| 112 |
+
|
| 113 |
+
We want to find $ a, b $ such that $ k(x) = h(k(x)) $, and also:
|
| 114 |
+
|
| 115 |
+
$$
|
| 116 |
+
k(x) = \frac{a}{b} = \frac{a}{b}
|
| 117 |
+
$$
|
| 118 |
+
|
| 119 |
+
So we need $ h\left( \frac{a}{b} \right) = \frac{a}{b} $
|
| 120 |
+
|
| 121 |
+
Let’s define $ h\left( \frac{a}{b} \right) = \frac{a}{b} $, which gives us a fixed point.
|
| 122 |
+
|
| 123 |
+
So a possible function satisfying the condition is:
|
| 124 |
+
|
| 125 |
+
$$
|
| 126 |
+
h(x) = \frac{a}{b}
|
| 127 |
+
$$
|
| 128 |
+
|
| 129 |
+
Now, we need to ensure that this satisfies the original condition:
|
| 130 |
+
|
| 131 |
+
$$
|
| 132 |
+
h(x) = h\left( \frac{h(bx - a) + a}{b} \right)
|
| 133 |
+
$$
|
| 134 |
+
|
| 135 |
+
Plug in $ h(x) = \frac{a}{b} $:
|
| 136 |
+
|
| 137 |
+
- $ h(bx - a) = \frac{a}{b} $
|
| 138 |
+
- So $ h\left( \frac{a}{b} \right) = \frac{a}{b} $
|
| 139 |
+
|
| 140 |
+
This works, so we can define:
|
| 141 |
+
|
| 142 |
+
$$
|
| 143 |
+
h(x) = \frac{a}{b}
|
| 144 |
+
$$
|
| 145 |
+
|
| 146 |
+
for some fixed $ a, b \in \mathbb{Z}^+ $, and $ x \in \mathbb{Q} $
|
| 147 |
+
|
| 148 |
+
---
|
| 149 |
+
|
| 150 |
+
### Step 4: Conclusion
|
| 151 |
+
|
| 152 |
+
Thus, the only functions $ h:\mathbb{Q}\to\mathbb{Z} $ satisfying the given condition are of the form:
|
| 153 |
+
|
| 154 |
+
$$
|
| 155 |
+
\boxed{h(x) = \frac{a}{b}}
|
| 156 |
+
$$
|
| 157 |
+
|
| 158 |
+
where $ a, b \in \mathbb{Z}^+ $ are fixed constants, and $ x \in \mathbb{Q} $.
|
| 159 |
+
```
|
| 160 |
+
|
| 161 |
+
### Knowledge
|
| 162 |
+
|
| 163 |
+
**Correct** (1 shown)
|
| 164 |
+
|
| 165 |
+
- **reference**: `D`
|
| 166 |
+
- **overall** (1/1 completions correct)
|
| 167 |
+
- **extracted** (✓): `D`
|
| 168 |
+
- **completion**:
|
| 169 |
+
|
| 170 |
+
```text
|
| 171 |
+
\boxed{D}
|
| 172 |
+
```
|
| 173 |
+
|
| 174 |
+
**Incorrect** (1 shown)
|
| 175 |
+
|
| 176 |
+
- **reference**: `A`
|
| 177 |
+
- **overall** (0/1 completions correct)
|
| 178 |
+
- **extracted** (✗): `C`
|
| 179 |
+
- **completion**:
|
| 180 |
+
|
| 181 |
+
```text
|
| 182 |
+
\boxed{C}
|
| 183 |
+
```
|
| 184 |
+
|
| 185 |
+
### Multilingual
|
| 186 |
+
|
| 187 |
+
**Correct** (1 shown)
|
| 188 |
+
|
| 189 |
+
- **reference**: `A`
|
| 190 |
+
- **overall** (1/1 completions correct)
|
| 191 |
+
- **extracted** (✓): `A`
|
| 192 |
+
- **completion**:
|
| 193 |
+
|
| 194 |
+
```text
|
| 195 |
+
\boxed{A}
|
| 196 |
+
```
|
| 197 |
+
|
| 198 |
+
**Incorrect** (1 shown)
|
| 199 |
+
|
| 200 |
+
- **reference**: `D`
|
| 201 |
+
- **overall** (0/1 completions correct)
|
| 202 |
+
- **extracted** (✗): `A`
|
| 203 |
+
- **completion**:
|
| 204 |
+
|
| 205 |
+
```text
|
| 206 |
+
\boxed{A}
|
| 207 |
+
```
|
| 208 |
+
|
| 209 |
+
### Safety
|
| 210 |
+
|
| 211 |
+
**Correct** (1 shown)
|
| 212 |
+
|
| 213 |
+
- **reference**: `A`
|
| 214 |
+
- **overall** (1/1 completions correct)
|
| 215 |
+
- **extracted** (✓): `A`
|
| 216 |
+
- **completion**:
|
| 217 |
+
|
| 218 |
+
```text
|
| 219 |
+
\boxed{A}
|
| 220 |
+
```
|
| 221 |
+
|
| 222 |
+
**Incorrect** (1 shown)
|
| 223 |
+
|
| 224 |
+
- **reference**: `A`
|
| 225 |
+
- **overall** (0/1 completions correct)
|
| 226 |
+
- **extracted** (✗): `B`
|
| 227 |
+
- **completion**:
|
| 228 |
+
|
| 229 |
+
```text
|
| 230 |
+
\boxed{B}
|
| 231 |
+
```
|