Upload README.md with huggingface_hub
Browse files
README.md
CHANGED
|
@@ -1,3 +1,54 @@
|
|
| 1 |
-
---
|
| 2 |
-
|
| 3 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
pipeline_tag: text-generation
|
| 3 |
+
library_name: transformers
|
| 4 |
+
tags:
|
| 5 |
+
- mining
|
| 6 |
+
- fp8
|
| 7 |
+
license: apache-2.0
|
| 8 |
+
language:
|
| 9 |
+
- ru
|
| 10 |
+
base_model: nn-tech/MetalGPT-1
|
| 11 |
+
---
|
| 12 |
+
|
| 13 |
+
## Description
|
| 14 |
+
|
| 15 |
+
**MetalGPT-1** is a model built upon the Qwen/Qwen3-32B and incorporates both continual pre-training and supervised fine-tuning on domain-specific data from the mining and metallurgy industry.
|
| 16 |
+
|
| 17 |
+
---
|
| 18 |
+
|
| 19 |
+
### Quantization
|
| 20 |
+
|
| 21 |
+
For convenience and improved performance, we also provide this FP8 checkpoint of the nn-tech/MetalGPT-1 model. Using FP8 precision enables faster inference and lower memory usage, while preserving model quality and numerical stability.
|
| 22 |
+
|
| 23 |
+
---
|
| 24 |
+
|
| 25 |
+
### VLLM usage
|
| 26 |
+
|
| 27 |
+
```bash
|
| 28 |
+
vllm serve nn-tech/MetalGPT-1-FP8 --reasoning-parser qwen3
|
| 29 |
+
|
| 30 |
+
```
|
| 31 |
+
|
| 32 |
+
```python
|
| 33 |
+
|
| 34 |
+
from openai import OpenAI
|
| 35 |
+
|
| 36 |
+
client = OpenAI(
|
| 37 |
+
base_url="http://localhost:8000/v1",
|
| 38 |
+
api_key="dummy"
|
| 39 |
+
)
|
| 40 |
+
|
| 41 |
+
response = client.chat.completions.create(
|
| 42 |
+
model="nn-tech/MetalGPT-1-FP8",
|
| 43 |
+
messages=[
|
| 44 |
+
{"role": "system", "content": "Ты специалист в области металлургии."},
|
| 45 |
+
{"role": "user", "content": "Назови плюсы и минусы хлоридной и сульфатной технологии производства никеля."}
|
| 46 |
+
],
|
| 47 |
+
temperature=0.7,
|
| 48 |
+
max_tokens=1024
|
| 49 |
+
)
|
| 50 |
+
|
| 51 |
+
print(response.choices[0].message.content)
|
| 52 |
+
|
| 53 |
+
```
|
| 54 |
+
|