Other
Transformers
Safetensors
PyTorch
English
vision-language-action
humanoid-robotics
telepathy
multimodal
robotics-control
lora
Instructions to use Veltraxor/Sigma with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use Veltraxor/Sigma with Transformers:
# Load model directly from transformers import AutoModel model = AutoModel.from_pretrained("Veltraxor/Sigma", dtype="auto") - Notebooks
- Google Colab
- Kaggle
Upload train_args.json
Browse files
storage/sigma_lora_out/train_args.json
CHANGED
|
@@ -28,14 +28,11 @@
|
|
| 28 |
"alpha_c": 1.0,
|
| 29 |
"lambda_sem": 1.2,
|
| 30 |
"lambda_intent": 0.8,
|
| 31 |
-
"lambda_tau": 0.
|
| 32 |
"beta_mi": 0.1,
|
| 33 |
"eta_var": 0.1,
|
| 34 |
"hard_mining_ratio": 0.3,
|
| 35 |
"hard_mining_lambda": 1.0,
|
| 36 |
-
"lambda_improve": 2.0,
|
| 37 |
-
"improve_margin": 0.01,
|
| 38 |
-
"improve_on_hard_only": true,
|
| 39 |
"loss_warmup_ratio": 0.4,
|
| 40 |
"lambda_sem_start": 0.2,
|
| 41 |
"lambda_intent_start": 0.2,
|
|
|
|
| 28 |
"alpha_c": 1.0,
|
| 29 |
"lambda_sem": 1.2,
|
| 30 |
"lambda_intent": 0.8,
|
| 31 |
+
"lambda_tau": 0.03,
|
| 32 |
"beta_mi": 0.1,
|
| 33 |
"eta_var": 0.1,
|
| 34 |
"hard_mining_ratio": 0.3,
|
| 35 |
"hard_mining_lambda": 1.0,
|
|
|
|
|
|
|
|
|
|
| 36 |
"loss_warmup_ratio": 0.4,
|
| 37 |
"lambda_sem_start": 0.2,
|
| 38 |
"lambda_intent_start": 0.2,
|