File size: 5,450 Bytes
ce61544
 
 
 
 
 
 
b09f138
 
 
ce61544
 
 
 
 
b09f138
 
 
ce61544
 
8d25b68
ce61544
 
 
 
 
 
8d25b68
ce61544
 
 
 
 
 
 
 
 
 
 
8d25b68
ce61544
6a9905a
ce61544
6a9905a
 
 
 
 
 
 
 
 
 
 
d034c0d
6a9905a
 
8d25b68
6a9905a
d034c0d
6a9905a
 
 
 
 
d034c0d
6a9905a
 
8d25b68
6a9905a
d034c0d
6a9905a
 
 
 
 
 
d034c0d
 
 
 
 
 
 
 
6a9905a
 
 
 
 
d034c0d
6a9905a
 
 
 
 
 
8d25b68
6a9905a
 
 
 
 
8d25b68
6a9905a
 
 
 
 
 
 
ce61544
 
6a9905a
8d25b68
6a9905a
 
 
 
 
8d25b68
6a9905a
8d25b68
6a9905a
 
 
 
ce61544
 
 
 
8d25b68
ce61544
 
 
 
 
 
 
b09f138
8d25b68
b09f138
 
 
 
 
ce61544
8d25b68
6a9905a
8d25b68
ce61544
b09f138
 
6a9905a
 
 
 
 
 
 
 
b09f138
ce61544
 
6a9905a
 
ce61544
8d25b68
6a9905a
ce61544
 
8d25b68
ce61544
 
 
 
 
6a9905a
8d25b68
6a9905a
 
 
 
ce61544
 
 
b09f138
8d25b68
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
---
viewer: false
tags: [uv-script, vllm, gpu, inference]
---

# vLLM Inference Scripts

Ready-to-run UV scripts for GPU-accelerated inference using [vLLM](https://github.com/vllm-project/vllm).

These scripts use [UV's inline script metadata](https://docs.astral.sh/uv/guides/scripts/) to automatically manage dependencies - just run with `uv run` and everything installs automatically!

## πŸ“‹ Available Scripts

### classify-dataset.py

Batch text classification using BERT-style encoder models (e.g., BERT, RoBERTa, DeBERTa, ModernBERT) with vLLM's optimized inference engine.

**Note**: This script is specifically for encoder-only classification models, not generative LLMs.

**Features:**

- πŸš€ High-throughput batch processing
- 🏷️ Automatic label mapping from model config
- πŸ“Š Confidence scores for predictions
- πŸ€— Direct integration with Hugging Face Hub

**Usage:**

```bash
# Local execution (requires GPU)
uv run classify-dataset.py \
    davanstrien/ModernBERT-base-is-new-arxiv-dataset \
    username/input-dataset \
    username/output-dataset \
    --inference-column text \
    --batch-size 10000
```

**HF Jobs execution:**

```bash
hf jobs uv run \
    --flavor l4x1 \
    --image vllm/vllm-openai \
    https://huggingface.co/datasets/uv-scripts/vllm/resolve/main/classify-dataset.py \
    davanstrien/ModernBERT-base-is-new-arxiv-dataset \
    username/input-dataset \
    username/output-dataset \
    --inference-column text \
    --batch-size 100000
```

### generate-responses.py

Generate responses for prompts using generative LLMs (e.g., Llama, Qwen, Mistral) with vLLM's high-performance inference engine.

**Features:**

- πŸ’¬ Automatic chat template application
- πŸ“ Support for both chat messages and plain text prompts
- πŸ”€ Multi-GPU tensor parallelism support
- πŸ“ Smart filtering for prompts exceeding context length
- πŸ“Š Comprehensive dataset cards with generation metadata
- ⚑ HF Transfer enabled for fast model downloads
- πŸŽ›οΈ Full control over sampling parameters
- 🎯 Sample limiting with `--max-samples` for testing

**Usage:**

```bash
# With chat-formatted messages (default)
uv run generate-responses.py \
    username/input-dataset \
    username/output-dataset \
    --messages-column messages \
    --max-tokens 1024

# With plain text prompts (NEW!)
uv run generate-responses.py \
    username/input-dataset \
    username/output-dataset \
    --prompt-column question \
    --max-tokens 1024 \
    --max-samples 100

# With custom model and parameters
uv run generate-responses.py \
    username/input-dataset \
    username/output-dataset \
    --model-id meta-llama/Llama-3.1-8B-Instruct \
    --prompt-column text \
    --temperature 0.9 \
    --top-p 0.95 \
    --max-model-len 8192
```

**HF Jobs execution (multi-GPU):**

```bash
hf jobs uv run \
    --flavor l4x4 \
    --image vllm/vllm-openai \
    -e UV_PRERELEASE=if-necessary \
    -s HF_TOKEN=hf_*** \
    https://huggingface.co/datasets/uv-scripts/vllm/raw/main/generate-responses.py \
    davanstrien/cards_with_prompts \
    davanstrien/test-generated-responses \
    --model-id Qwen/Qwen3-30B-A3B-Instruct-2507 \
    --gpu-memory-utilization 0.9 \
    --max-tokens 600 \
    --max-model-len 8000
```

### Multi-GPU Tensor Parallelism

- Auto-detects available GPUs by default
- Use `--tensor-parallel-size` to manually specify
- Required for models larger than single GPU memory (e.g., 30B+ models)

### Handling Long Contexts

The generate-responses.py script includes smart prompt filtering:

- **Default behavior**: Skips prompts exceeding max_model_len
- **Use `--max-model-len`**: Limit context to reduce memory usage
- **Use `--no-skip-long-prompts`**: Fail on long prompts instead of skipping
- Skipped prompts receive empty responses and are logged

## πŸ“š About vLLM

vLLM is a high-throughput inference engine optimized for:

- Fast model serving with PagedAttention
- Efficient batch processing
- Support for various model architectures
- Seamless integration with Hugging Face models

## πŸ”§ Technical Details

### UV Script Benefits

- **Zero setup**: Dependencies install automatically on first run
- **Reproducible**: Locked dependencies ensure consistent behavior
- **Self-contained**: Everything needed is in the script file
- **Direct execution**: Run from local files or URLs

### Dependencies

Scripts use UV's inline metadata for automatic dependency management:

```python
# /// script
# requires-python = ">=3.10"
# dependencies = [
#     "datasets",
#     "flashinfer-python",
#     "huggingface-hub[hf_transfer]",
#     "torch",
#     "transformers",
#     "vllm",
# ]
# ///
```

For bleeding-edge features, use the `UV_PRERELEASE=if-necessary` environment variable to allow pre-release versions when needed.

### Docker Image

For HF Jobs, we recommend the official vLLM Docker image: `vllm/vllm-openai`

This image includes:

- Pre-installed CUDA libraries
- vLLM and all dependencies
- UV package manager
- Optimized for GPU inference

### Environment Variables

- `HF_TOKEN`: Your Hugging Face authentication token (auto-detected if logged in)
- `UV_PRERELEASE=if-necessary`: Allow pre-release packages when required
- `HF_HUB_ENABLE_HF_TRANSFER=1`: Automatically enabled for faster downloads

## πŸ”— Resources

- [vLLM Documentation](https://docs.vllm.ai/)
- [UV Documentation](https://docs.astral.sh/uv/)
- [UV Scripts Organization](https://huggingface.co/uv-scripts)