Commit
Β·
6a9905a
1
Parent(s):
68dd0ca
Update README.md with enhanced usage instructions for classify-dataset.py and generate-responses.py, including multi-GPU support and environment variable details.
Browse files
README.md
CHANGED
@@ -36,20 +36,62 @@ uv run classify-dataset.py \
|
|
36 |
|
37 |
**HF Jobs execution:**
|
38 |
```bash
|
39 |
-
|
40 |
--flavor l4x1 \
|
41 |
-
--
|
42 |
-
vllm/
|
43 |
-
/
|
44 |
-
|
45 |
-
|
46 |
-
|
47 |
-
|
48 |
-
|
49 |
-
|
50 |
-
|
51 |
-
|
52 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
53 |
```
|
54 |
|
55 |
## π― Requirements
|
@@ -62,13 +104,27 @@ All scripts in this collection require:
|
|
62 |
## π Performance Tips
|
63 |
|
64 |
### GPU Selection
|
65 |
-
- **L4 GPU** (`--flavor l4x1`): Best value for classification
|
66 |
-
- **
|
67 |
-
-
|
|
|
|
|
68 |
|
69 |
### Batch Sizes
|
70 |
-
- **
|
71 |
-
- **
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
72 |
|
73 |
## π About vLLM
|
74 |
|
@@ -87,22 +143,25 @@ vLLM is a high-throughput inference engine optimized for:
|
|
87 |
- **Direct execution**: Run from local files or URLs
|
88 |
|
89 |
### Dependencies
|
90 |
-
Scripts use UV's inline metadata
|
91 |
```python
|
92 |
# /// script
|
93 |
# requires-python = ">=3.10"
|
94 |
-
# dependencies = [
|
95 |
-
#
|
96 |
-
#
|
97 |
-
#
|
98 |
-
#
|
99 |
-
#
|
100 |
-
#
|
|
|
101 |
# ///
|
102 |
```
|
103 |
|
|
|
|
|
104 |
### Docker Image
|
105 |
-
For HF Jobs, we
|
106 |
|
107 |
This image includes:
|
108 |
- Pre-installed CUDA libraries
|
@@ -110,6 +169,11 @@ This image includes:
|
|
110 |
- UV package manager
|
111 |
- Optimized for GPU inference
|
112 |
|
|
|
|
|
|
|
|
|
|
|
113 |
## π Contributing
|
114 |
|
115 |
Have a vLLM script to share? We welcome contributions that:
|
|
|
36 |
|
37 |
**HF Jobs execution:**
|
38 |
```bash
|
39 |
+
hf jobs uv run \
|
40 |
--flavor l4x1 \
|
41 |
+
--image vllm/vllm-openai \
|
42 |
+
https://huggingface.co/datasets/uv-scripts/vllm/resolve/main/classify-dataset.py \
|
43 |
+
davanstrien/ModernBERT-base-is-new-arxiv-dataset \
|
44 |
+
username/input-dataset \
|
45 |
+
username/output-dataset \
|
46 |
+
--inference-column text \
|
47 |
+
--batch-size 100000
|
48 |
+
```
|
49 |
+
|
50 |
+
### generate-responses.py
|
51 |
+
|
52 |
+
Generate responses for chat-formatted prompts using generative LLMs (e.g., Llama, Qwen, Mistral) with vLLM's high-performance inference engine.
|
53 |
+
|
54 |
+
**Features:**
|
55 |
+
- π¬ Automatic chat template application
|
56 |
+
- π Multi-GPU tensor parallelism support
|
57 |
+
- π Smart filtering for prompts exceeding context length
|
58 |
+
- π Comprehensive dataset cards with generation metadata
|
59 |
+
- β‘ HF Transfer enabled for fast model downloads
|
60 |
+
- ποΈ Full control over sampling parameters
|
61 |
+
|
62 |
+
**Usage:**
|
63 |
+
```bash
|
64 |
+
# Local execution with default Qwen model
|
65 |
+
uv run generate-responses.py \
|
66 |
+
username/input-dataset \
|
67 |
+
username/output-dataset \
|
68 |
+
--messages-column messages \
|
69 |
+
--max-tokens 1024
|
70 |
+
|
71 |
+
# With custom model and parameters
|
72 |
+
uv run generate-responses.py \
|
73 |
+
username/input-dataset \
|
74 |
+
username/output-dataset \
|
75 |
+
--model-id meta-llama/Llama-3.1-8B-Instruct \
|
76 |
+
--temperature 0.9 \
|
77 |
+
--top-p 0.95 \
|
78 |
+
--max-model-len 8192
|
79 |
+
```
|
80 |
+
|
81 |
+
**HF Jobs execution (multi-GPU):**
|
82 |
+
```bash
|
83 |
+
hf jobs uv run \
|
84 |
+
--flavor l4x4 \
|
85 |
+
--image vllm/vllm-openai \
|
86 |
+
-e UV_PRERELEASE=if-necessary \
|
87 |
+
-e HF_TOKEN=hf_*** \
|
88 |
+
https://huggingface.co/datasets/uv-scripts/vllm/raw/main/generate-responses.py \
|
89 |
+
davanstrien/cards_with_prompts \
|
90 |
+
davanstrien/test-generated-responses \
|
91 |
+
--model-id Qwen/Qwen3-30B-A3B-Instruct-2507 \
|
92 |
+
--gpu-memory-utilization 0.9 \
|
93 |
+
--max-tokens 600 \
|
94 |
+
--max-model-len 8000
|
95 |
```
|
96 |
|
97 |
## π― Requirements
|
|
|
104 |
## π Performance Tips
|
105 |
|
106 |
### GPU Selection
|
107 |
+
- **L4 GPU** (`--flavor l4x1`): Best value for classification and smaller models
|
108 |
+
- **L4x4** (`--flavor l4x4`): Multi-GPU setup for large models (30B+ parameters)
|
109 |
+
- **A10 GPU** (`--flavor a10g-large`): Higher memory for larger models
|
110 |
+
- **A100** (`--flavor a100-large`): Maximum performance for demanding workloads
|
111 |
+
- Adjust batch size and tensor parallelism based on GPU configuration
|
112 |
|
113 |
### Batch Sizes
|
114 |
+
- **Classification**: Start with 10,000 locally, up to 100,000 on HF Jobs
|
115 |
+
- **Generation**: vLLM handles batching automatically - no manual configuration needed
|
116 |
+
|
117 |
+
### Multi-GPU Tensor Parallelism
|
118 |
+
- Auto-detects available GPUs by default
|
119 |
+
- Use `--tensor-parallel-size` to manually specify
|
120 |
+
- Required for models larger than single GPU memory (e.g., 30B+ models)
|
121 |
+
|
122 |
+
### Handling Long Contexts
|
123 |
+
The generate-responses.py script includes smart prompt filtering:
|
124 |
+
- **Default behavior**: Skips prompts exceeding max_model_len
|
125 |
+
- **Use `--max-model-len`**: Limit context to reduce memory usage
|
126 |
+
- **Use `--no-skip-long-prompts`**: Fail on long prompts instead of skipping
|
127 |
+
- Skipped prompts receive empty responses and are logged
|
128 |
|
129 |
## π About vLLM
|
130 |
|
|
|
143 |
- **Direct execution**: Run from local files or URLs
|
144 |
|
145 |
### Dependencies
|
146 |
+
Scripts use UV's inline metadata for automatic dependency management:
|
147 |
```python
|
148 |
# /// script
|
149 |
# requires-python = ">=3.10"
|
150 |
+
# dependencies = [
|
151 |
+
# "datasets",
|
152 |
+
# "flashinfer-python",
|
153 |
+
# "huggingface-hub[hf_transfer]",
|
154 |
+
# "torch",
|
155 |
+
# "transformers",
|
156 |
+
# "vllm",
|
157 |
+
# ]
|
158 |
# ///
|
159 |
```
|
160 |
|
161 |
+
For bleeding-edge features, use the `UV_PRERELEASE=if-necessary` environment variable to allow pre-release versions when needed.
|
162 |
+
|
163 |
### Docker Image
|
164 |
+
For HF Jobs, we recommend the official vLLM Docker image: `vllm/vllm-openai`
|
165 |
|
166 |
This image includes:
|
167 |
- Pre-installed CUDA libraries
|
|
|
169 |
- UV package manager
|
170 |
- Optimized for GPU inference
|
171 |
|
172 |
+
### Environment Variables
|
173 |
+
- `HF_TOKEN`: Your Hugging Face authentication token (auto-detected if logged in)
|
174 |
+
- `UV_PRERELEASE=if-necessary`: Allow pre-release packages when required
|
175 |
+
- `HF_HUB_ENABLE_HF_TRANSFER=1`: Automatically enabled for faster downloads
|
176 |
+
|
177 |
## π Contributing
|
178 |
|
179 |
Have a vLLM script to share? We welcome contributions that:
|