Commit
•
0c39999
1
Parent(s):
1eb4934
Update README.md
Browse files
README.md
CHANGED
|
@@ -23,6 +23,8 @@ Try the 3.8B model here: [Playground](https://huggingface.co/spaces/numind/NuExt
|
|
| 23 |
|
| 24 |
We also provide a tiny (0.5B) version which is based on Qwen2.5-0.5B: [NuExtract-tiny-v1.5](https://huggingface.co/numind/NuExtract-tiny-v1.5)
|
| 25 |
|
|
|
|
|
|
|
| 26 |
## Benchmark
|
| 27 |
|
| 28 |
Zero-shot performance (English):
|
|
|
|
| 23 |
|
| 24 |
We also provide a tiny (0.5B) version which is based on Qwen2.5-0.5B: [NuExtract-tiny-v1.5](https://huggingface.co/numind/NuExtract-tiny-v1.5)
|
| 25 |
|
| 26 |
+
⚠️ We recommend using NuExtract with a temperature at or very close to 0. Some inference frameworks, such as Ollama, use a default of 0.7 which is not well suited to pure extraction tasks.
|
| 27 |
+
|
| 28 |
## Benchmark
|
| 29 |
|
| 30 |
Zero-shot performance (English):
|