Instructions to use PrunaAI/WeiboAI-VibeThinker-3B-HQQ-4bit-smashed with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Pruna AI
How to use PrunaAI/WeiboAI-VibeThinker-3B-HQQ-4bit-smashed with Pruna AI:
from pruna import PrunaModel model = PrunaModel.from_pretrained("PrunaAI/WeiboAI-VibeThinker-3B-HQQ-4bit-smashed") - Notebooks
- Google Colab
- Kaggle
| { | |
| "batcher": null, | |
| "cacher": null, | |
| "compiler": null, | |
| "factorizer": null, | |
| "pruner": null, | |
| "quantizer": "hqq", | |
| "hqq_backend": "torchao_int4", | |
| "hqq_compute_dtype": "torch.float16", | |
| "hqq_group_size": 64, | |
| "hqq_weight_bits": 4, | |
| "batch_size": 1, | |
| "device": "cuda", | |
| "save_fns": [ | |
| "hqq" | |
| ], | |
| "load_fns": [ | |
| "hqq" | |
| ], | |
| "reapply_after_load": { | |
| "factorizer": null, | |
| "pruner": null, | |
| "quantizer": null, | |
| "cacher": null, | |
| "compiler": null, | |
| "batcher": null | |
| }, | |
| "api_key": null | |
| } |