Instructions to use array/sat-dynamic-13b with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use array/sat-dynamic-13b with Transformers:
# Load model directly from transformers import AutoModel model = AutoModel.from_pretrained("array/sat-dynamic-13b", device_map="auto") - Notebooks
- Google Colab
- Kaggle
Download adapter_config.json from array/sat-dynamic-13b: direct link, hf CLI and curl.
- Browser
- Download file 720 Bytes
-
https://huggingface.co/array/sat-dynamic-13b/resolve/main/adapter_config.json
- Command line
-
hf download hf://array/sat-dynamic-13b/adapter_config.json
-
curl -L -o adapter_config.json https://huggingface.co/array/sat-dynamic-13b/resolve/main/adapter_config.json
720 Bytes
| { | |
| "alpha_pattern": {}, | |
| "auto_mapping": { | |
| "base_model_class": "LlavaModel_13B_Interface", | |
| "parent_library": "models.model_interface" | |
| }, | |
| "base_model_name_or_path": null, | |
| "bias": "none", | |
| "fan_in_fan_out": false, | |
| "inference_mode": true, | |
| "init_lora_weights": true, | |
| "layer_replication": null, | |
| "layers_pattern": null, | |
| "layers_to_transform": null, | |
| "loftq_config": {}, | |
| "lora_alpha": 512, | |
| "lora_dropout": 0.1, | |
| "megatron_config": null, | |
| "megatron_core": "megatron.core", | |
| "modules_to_save": [], | |
| "peft_type": "LORA", | |
| "r": 256, | |
| "rank_pattern": {}, | |
| "revision": null, | |
| "target_modules": [ | |
| "q_proj", | |
| "v_proj" | |
| ], | |
| "task_type": null, | |
| "use_dora": false, | |
| "use_rslora": false | |
| } |