Instructions to use zzyzyzy/ReconVLA with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use zzyzyzy/ReconVLA with Transformers:
# Load model directly from transformers import AutoModel model = AutoModel.from_pretrained("zzyzyzy/ReconVLA", dtype="auto") - Notebooks
- Google Colab
- Kaggle
Update checkpoint-5554/config.json
Browse files
checkpoint-5554/config.json
CHANGED
|
@@ -40,7 +40,7 @@
|
|
| 40 |
"mm_vision_select_layer": -2,
|
| 41 |
"mm_vision_tower": "./siglip-so400m-patch14-384",
|
| 42 |
"mm_vision_tower_lr": null,
|
| 43 |
-
"model_type": "
|
| 44 |
"num_attention_heads": 28,
|
| 45 |
"num_hidden_layers": 28,
|
| 46 |
"num_key_value_heads": 4,
|
|
|
|
| 40 |
"mm_vision_select_layer": -2,
|
| 41 |
"mm_vision_tower": "./siglip-so400m-patch14-384",
|
| 42 |
"mm_vision_tower_lr": null,
|
| 43 |
+
"model_type": "recon_qwen2",
|
| 44 |
"num_attention_heads": 28,
|
| 45 |
"num_hidden_layers": 28,
|
| 46 |
"num_key_value_heads": 4,
|