Instructions to use sinequa/vectorizer.vanilla with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use sinequa/vectorizer.vanilla with Transformers:
# Load model directly from transformers import AutoTokenizer, AutoModel tokenizer = AutoTokenizer.from_pretrained("sinequa/vectorizer.vanilla") model = AutoModel.from_pretrained("sinequa/vectorizer.vanilla", device_map="auto") - Notebooks
- Google Colab
- Kaggle
Update minimum CUDA version
Browse files
README.md
CHANGED
|
@@ -39,7 +39,7 @@ Note that the relevance score is computed as an average over 14 retrieval datase
|
|
| 39 |
| NVIDIA L4 | FP16 | 1 ms | 5 ms |
|
| 40 |
| NVIDIA L4 | FP32 | 3 ms | 25 ms |
|
| 41 |
|
| 42 |
-
##
|
| 43 |
|
| 44 |
| Quantization type | Memory |
|
| 45 |
|:-------------------------------------------------|-----------:|
|
|
@@ -53,7 +53,7 @@ can be around 0.5 to 1 GiB depending on the used GPU.
|
|
| 53 |
## Requirements
|
| 54 |
|
| 55 |
- Minimal Sinequa version: 11.10.0
|
| 56 |
-
- [
|
| 57 |
|
| 58 |
## Model Details
|
| 59 |
|
|
|
|
| 39 |
| NVIDIA L4 | FP16 | 1 ms | 5 ms |
|
| 40 |
| NVIDIA L4 | FP32 | 3 ms | 25 ms |
|
| 41 |
|
| 42 |
+
## GPU Memory usage
|
| 43 |
|
| 44 |
| Quantization type | Memory |
|
| 45 |
|:-------------------------------------------------|-----------:|
|
|
|
|
| 53 |
## Requirements
|
| 54 |
|
| 55 |
- Minimal Sinequa version: 11.10.0
|
| 56 |
+
- [CUDA compute capability](https://developer.nvidia.com/cuda-gpus): above 7.5
|
| 57 |
|
| 58 |
## Model Details
|
| 59 |
|