Instructions to use coded190/pingti_palm_openvla with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- PEFT
How to use coded190/pingti_palm_openvla with PEFT:
from peft import PeftModel from transformers import AutoModelForCausalLM base_model = AutoModelForCausalLM.from_pretrained("openvla/openvla-7b") model = PeftModel.from_pretrained(base_model, "coded190/pingti_palm_openvla") - Notebooks
- Google Colab
- Kaggle
File size: 427 Bytes
12b2447 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 | {
"vla_path": "openvla/openvla-7b",
"data_jsonl": null,
"lerobot_repo_ids": [
"coded190/pingti_palm_tree",
"coded190/isaac_so_arm101_vla_v2"
],
"image_root": null,
"output_dir": "outputs/openvla_lora",
"batch_size": 16,
"grad_accum_steps": 1,
"max_steps": 2500,
"save_steps": 200,
"learning_rate": 0.0005,
"mixed_precision": "bf16",
"action_dim": 7,
"predict_stop_token": true,
"seed": 0
} |