| """ |
| run_libero_eval.py |
| |
| Runs a model in a LIBERO simulation environment. |
| |
| Usage: |
| # OpenVLA: |
| # IMPORTANT: Set `center_crop=True` if model is fine-tuned with augmentations |
| oc slot |
| CUDA_VISIBLE_DEVICES=0 python run_oc_model_sample.py \ |
| --model_family openvla \ |
| --saved_dir output_oc |
| """ |
|
|
| import random |
| import os |
| import sys |
| from dataclasses import dataclass |
| from pathlib import Path |
| from typing import Optional, Union |
| import torch |
| import draccus |
| import numpy as np |
| import tqdm |
| from transformers import AutoConfig, AutoImageProcessor, AutoModelForVision2Seq, AutoProcessor |
|
|
| |
|
|
| |
| |
| |
| |
| |
| |
| from robot_utils import ( |
| DATE_TIME, |
| ) |
|
|
| from prismatic.extern.hf.configuration_prismatic import OpenVLAConfig |
| from prismatic.extern.hf.modeling_prismatic import OpenVLAForActionPrediction, EmbodiedObjectSlot, EmbodiedRelationSlot, EmbodiedObject_LangSlot, EmbodiedRelation_LangSlot |
| from prismatic.extern.hf.processing_prismatic import PrismaticImageProcessor, PrismaticProcessor |
| from peft import PeftModel |
| from safetensors.torch import load_file |
|
|
| import json |
|
|
| def get_processor(cfg): |
| """Get VLA model's Hugging Face processor.""" |
| processor = AutoProcessor.from_pretrained(os.path.dirname(cfg.custom_param_checkpoint), trust_remote_code=False) |
| return processor |
|
|
| def get_vla(cfg): |
| """Loads and returns a VLA model from checkpoint.""" |
| |
| print("[*] Instantiating Pretrained VLA model") |
| |
| print("[*] Loading in BF16 with Flash-Attention Enabled") |
|
|
| AutoConfig.register("openvla", OpenVLAConfig) |
| AutoImageProcessor.register(OpenVLAConfig, PrismaticImageProcessor) |
| AutoProcessor.register(OpenVLAConfig, PrismaticProcessor) |
| AutoModelForVision2Seq.register(OpenVLAConfig, OpenVLAForActionPrediction) |
|
|
| base_vla = AutoModelForVision2Seq.from_pretrained( |
| 'output_hf_model_openx', |
| attn_implementation="flash_attention_2", |
| torch_dtype=torch.bfloat16, |
| |
| load_in_8bit=cfg.load_in_8bit, |
| load_in_4bit=cfg.load_in_4bit, |
| low_cpu_mem_usage=True, |
| trust_remote_code=False, |
| ).to('cuda') |
| print("Loaded base.") |
| |
| base_vla = PeftModel.from_pretrained(base_vla, cfg.pretrained_checkpoint) |
| base_vla = base_vla.merge_and_unload() |
| print("Merged CustomOpenVLA LLM LoRA.") |
|
|
| if cfg.slot_type == 'oc': |
| vla = EmbodiedObject_LangSlot(base_model=base_vla, number_of_slots = cfg.number_of_slots) |
| elif cfg.slot_type == 'orc': |
| vla = EmbodiedRelation_LangSlot(base_model=base_vla, number_of_slots = cfg.number_of_slots) |
| |
| weights = load_file(cfg.custom_param_checkpoint) |
| vla.load_state_dict(weights, strict=False) |
| vla.object_centric_tokenizer.requires_grad_(False) |
| vla.object_centric_bbox_head.requires_grad_(False) |
| vla.object_centric_mask_head.requires_grad_(False) |
| vla.requires_grad_(False) |
| vla = vla.to('cuda') |
|
|
| |
| |
| |
| if not cfg.load_in_8bit and not cfg.load_in_4bit: |
| vla = vla.to('cuda') |
|
|
| |
| dataset_statistics_path = os.path.join(os.path.dirname(cfg.custom_param_checkpoint), "dataset_statistics.json") |
| if os.path.isfile(dataset_statistics_path): |
| with open(dataset_statistics_path, "r") as f: |
| norm_stats = json.load(f) |
| vla.norm_stats = norm_stats |
| else: |
| print( |
| "WARNING: No local dataset_statistics.json file found for current checkpoint.\n" |
| "You can ignore this if you are loading the base VLA (i.e. not fine-tuned) checkpoint." |
| "Otherwise, you may run into errors when trying to call `predict_action()` due to an absent `unnorm_key`." |
| ) |
|
|
| return vla |
|
|
| def get_model(cfg, wrap_diffusion_policy_for_droid=False): |
| """Load model for evaluation.""" |
| if cfg.model_family == "openvla" or cfg.model_family == "customvla" or cfg.model_family == 'objectvla': |
| model = get_vla(cfg) |
| else: |
| raise ValueError("Unexpected `model_family` found in config.") |
| print(f"Loaded model: {type(model)}") |
| return model |
|
|
| def visualize_all(visuals): |
| |
| row1 = [ |
| visuals['exo_rgb'], |
| visuals['exo_rgb_boxed'], |
| visuals['exo_rgb_boxed_contact'], |
| visuals['exo_depth'], |
| visuals['exo_seg'], |
| ] |
|
|
| row2 = [ |
| visuals['ego_rgb'], |
| visuals['ego_rgb_boxed'], |
| visuals['ego_rgb_boxed_contact'], |
| visuals['ego_depth'], |
| visuals['ego_seg'], |
| ] |
|
|
| |
| def preprocess_image(image, target_size=(300, 300)): |
| if len(image.shape) == 2: |
| image = cv2.cvtColor(image, cv2.COLOR_GRAY2BGR) |
| image = cv2.cvtColor(image, cv2.COLOR_BGR2RGB) |
| image = cv2.resize(image, target_size) |
| return image |
|
|
| |
| row1_concat = np.hstack([preprocess_image(img) for img in row1]) |
| row2_concat = np.hstack([preprocess_image(img) for img in row2]) |
|
|
| |
| final_image = np.vstack([row1_concat, row2_concat]) |
| cv2.imshow('data', final_image) |
|
|
| from dataclasses import dataclass, field |
|
|
| def get_libero_dummy_action(model_family: str): |
| """Get dummy/no-op action, used to roll out the simulation while the robot does nothing.""" |
| return [0, 0, 0, 0, 0, 0, -1] |
|
|
| @dataclass |
| class GenerateConfig: |
| |
|
|
| |
| |
| |
| model_family: str = "openvla" |
| slot_type: str = "oc" |
| number_of_slots: int = 16 |
| pretrained_checkpoint: Union[str, Path] = "oc_model" |
| custom_param_checkpoint: Union[str, Path] = "oc_model/relate_object_bboxes_w_mask_actionable_s16h1.safetensors" |
| load_in_8bit: bool = False |
| load_in_4bit: bool = False |
| |
| center_crop: bool = False |
|
|
| |
| |
| |
| task_suite_name: str = "real_data" |
| num_steps_wait: int = 20 |
| num_trials_per_task: int = 20 |
|
|
| |
| |
| |
| run_id_note: Optional[str] = None |
| local_log_dir: str = "./experiments/logs" |
| saved_dir: str = "libero_goal_01_slots" |
|
|
| use_wandb: bool = False |
| wandb_project: str = "YOUR_WANDB_PROJECT" |
| wandb_entity: str = "YOUR_WANDB_ENTITY" |
|
|
| seed: int = 7 |
|
|
| writing_extra_waits: bool = False |
| |
|
|
| import re |
|
|
| def natural_key(s): |
| return [int(text) if text.isdigit() else text for text in re.split(r'(\d+)', s)] |
|
|
| import cv2 |
| import pickle |
|
|
| IMAGE_RESOLUTION = 256 |
| from PIL import Image |
|
|
| def process_inputs(processor, pixel_values, device): |
| image = Image.fromarray(pixel_values) |
| image = image.convert("RGB") |
| |
| |
| action_dim = 7; future_horizon = 5 |
| prompt = f"In: What action should the robot take?\nOut: " |
| placeholder_seg = "_ _ _ _ _ _ _ _" |
| for fidx in range(future_horizon-1): |
| placeholder_seg += " _ _ _ _ _ _ _" |
| prompt = prompt + placeholder_seg + "</s>" |
|
|
| |
| inputs = processor(prompt, image).to(device, dtype=torch.float16) |
| inputs['input_ids'][0][-2-(action_dim*future_horizon)] = 29871 |
| |
| |
| |
| return inputs |
|
|
| def set_seed_everywhere(seed: int): |
| """Sets the random seed for Python, NumPy, and PyTorch functions.""" |
| torch.manual_seed(seed) |
| torch.cuda.manual_seed_all(seed) |
| np.random.seed(seed) |
| random.seed(seed) |
| torch.backends.cudnn.deterministic = True |
| torch.backends.cudnn.benchmark = False |
| os.environ["PYTHONHASHSEED"] = str(seed) |
|
|
|
|
| def get_image_resize_size(cfg): |
| """ |
| Gets image resize size for a model class. |
| If `resize_size` is an int, then the resized image will be a square. |
| Else, the image will be a rectangle. |
| """ |
| if cfg.model_family == "openvla" or cfg.model_family == "customvla" or cfg.model_family == 'objectvla': |
| resize_size = 224 |
| else: |
| raise ValueError("Unexpected `model_family` found in config.") |
| return resize_size |
|
|
|
|
| def get_current_slots(vla, batch, task_texts, in_past_tokens, device_id='cuda'): |
| pixel_values = batch["pixel_values"].to(torch.bfloat16).to(device_id) |
| outputs = vla.get_obj_slots(pixel_values, task_texts, in_past_tokens) |
| return outputs |
|
|
| def get_normalized_actions(vla, batch, slot_outputs, device_id): |
| |
| patch_features = slot_outputs['patch_features'] |
| slotted_features = slot_outputs['visual_tokens'] |
| if 'interactable_features' in slot_outputs: |
| slotted_features, _ = vla.select_top_k_slots( |
| slotted_features, slot_outputs['interactable_features'], k=4) |
| clip_embeddings = slot_outputs['texts'] |
| clip_attention_mask = torch.logical_not(slot_outputs['texts_attn']) |
| llama_input_ids = batch["input_ids"].to(device_id) |
| llama_attention_mask = batch["attention_mask"].to(device_id) |
|
|
| continuous_actions_pred = vla.decode_continuous_actions( |
| patch_features=patch_features, |
| slotted_features=slotted_features, |
| clip_embeddings=clip_embeddings, |
| clip_attention_mask=clip_attention_mask, |
| llama_input_ids=llama_input_ids, |
| llama_attention_mask=llama_attention_mask, |
| llama_labels=None |
| ) |
| action_chunk, action_dim = continuous_actions_pred.shape[1:] |
|
|
| return continuous_actions_pred[:,0,:] |
|
|
| def denormalize_actions(model, unnorm_key, normalized_actions): |
| |
| action_norm_stats = model.get_action_stats(unnorm_key) |
| mask = action_norm_stats.get("mask", np.ones_like(action_norm_stats["q01"], dtype=bool)) |
| action_high, action_low = np.array(action_norm_stats["q99"]), np.array(action_norm_stats["q01"]) |
| actions = np.where( |
| mask, |
| 0.5 * (normalized_actions + 1) * (action_high - action_low) + action_low, |
| normalized_actions, |
| ) |
| return actions |
|
|
| def normalize_gripper_action(action, binarize=True): |
| """ |
| Changes gripper action (last dimension of action vector) from [0,1] to [-1,+1]. |
| Necessary for some environments (not Bridge) because the dataset wrapper standardizes gripper actions to [0,1]. |
| Note that unlike the other action dimensions, the gripper action is not normalized to [-1,+1] by default by |
| the dataset wrapper. |
| |
| Normalization formula: y = 2 * (x - orig_low) / (orig_high - orig_low) - 1 |
| """ |
| |
| orig_low, orig_high = 0.0, 1.0 |
| action[..., -1] = 2 * (action[..., -1] - orig_low) / (orig_high - orig_low) - 1 |
|
|
| if binarize: |
| |
| action[..., -1] = np.sign(action[..., -1]) |
|
|
| return action |
|
|
| |
|
|
| @draccus.wrap() |
| def eval_libero(cfg: GenerateConfig) -> None: |
| print(cfg.pretrained_checkpoint) |
| assert cfg.pretrained_checkpoint is not None, "cfg.pretrained_checkpoint must not be None!" |
| |
| |
| assert not (cfg.load_in_8bit and cfg.load_in_4bit), "Cannot use both 8-bit and 4-bit quantization!" |
|
|
| |
| set_seed_everywhere(cfg.seed) |
|
|
| |
| cfg.unnorm_key = cfg.task_suite_name |
|
|
| |
| model = get_model(cfg) |
|
|
| |
| if cfg.model_family == "openvla": |
| |
| |
| if cfg.unnorm_key not in model.norm_stats and f"{cfg.unnorm_key}_no_noops" in model.norm_stats: |
| cfg.unnorm_key = f"{cfg.unnorm_key}_no_noops" |
| assert cfg.unnorm_key in model.norm_stats, f"Action un-norm key {cfg.unnorm_key} not found in VLA `norm_stats`!" |
|
|
| |
| processor = None |
| if cfg.model_family == "openvla": |
| processor = get_processor(cfg) |
|
|
| |
| run_id = f"EVAL-{cfg.task_suite_name}-{cfg.model_family}-{DATE_TIME}" |
| if cfg.run_id_note is not None: |
| run_id += f"--{cfg.run_id_note}" |
| os.makedirs(cfg.local_log_dir, exist_ok=True) |
| local_log_filepath = os.path.join(cfg.local_log_dir, run_id + ".txt") |
| log_file = open(local_log_filepath, "w") |
| print(f"Logging to local log file: {local_log_filepath}") |
|
|
| |
| benchmark_dict = benchmark.get_benchmark_dict() |
| task_suite = benchmark_dict[cfg.task_suite_name]() |
| num_tasks_in_suite = task_suite.n_tasks |
| print(f"Task suite: {cfg.task_suite_name}") |
| log_file.write(f"Task suite: {cfg.task_suite_name}\n") |
|
|
| |
| resize_size = get_image_resize_size(cfg) |
|
|
| |
| total_episodes, total_successes = 0, 0 |
| extra_waits_dict = {} |
| for task_id in tqdm.tqdm(range(num_tasks_in_suite)): |
| |
| task = task_suite.get_task(task_id) |
|
|
| |
| initial_states = task_suite.get_task_init_states(task_id) |
|
|
| |
| env, task_description = get_libero_env(task, cfg.model_family, resolution=256) |
|
|
| |
| task_episodes, task_successes = 0, 0 |
| extra_waits = [] |
| for episode_idx in tqdm.tqdm(range(cfg.num_trials_per_task)): |
| print(f"\nTask: {task_description}") |
| log_file.write(f"\nTask: {task_description}\n") |
|
|
| |
| env.reset() |
|
|
| |
| obs = env.set_init_state(initial_states[episode_idx]) |
| |
| |
| t = 0 |
| replay_images = [] |
| if cfg.task_suite_name == "libero_spatial": |
| max_steps = 220 |
| elif cfg.task_suite_name == "libero_object": |
| max_steps = 280 |
| elif cfg.task_suite_name == "libero_goal": |
| max_steps = 300 |
| elif cfg.task_suite_name == "libero_10": |
| max_steps = 520 |
| elif cfg.task_suite_name == "libero_90": |
| max_steps = 400 |
|
|
| t = 0 |
| while t < cfg.num_steps_wait: |
| |
| |
| |
| obs, reward, done, info = env.step(get_libero_dummy_action(cfg.model_family)) |
| t += 1 |
| |
| |
| task_texts = [task_description] |
| texts = ['robot ' + task_description] |
| img = get_libero_image(obs, resize_size) |
| img = img[:,::-1,:] |
| batch_data = process_inputs(processor=processor, pixel_values=img, device='cuda') |
|
|
| horizon = 1 |
| batch_data['pixel_values'] = batch_data['pixel_values'].unsqueeze(1).repeat(1,horizon,1,1,1) |
| |
| |
| with torch.autocast("cuda", dtype=torch.bfloat16): |
| slot_outputs = get_current_slots(model, batch_data, texts, in_past_tokens=None, device_id='cuda') |
| |
| |
| with torch.autocast("cuda", dtype=torch.bfloat16): |
| actions = get_normalized_actions(model, batch_data, slot_outputs, device_id='cuda') |
| actions = actions.detach().float().cpu().numpy() |
| unnormalized_actions = denormalize_actions(model, cfg.unnorm_key, actions) |
|
|
| |
| while t < max_steps + cfg.num_steps_wait: |
| |
| img = get_libero_image(obs, resize_size) |
| img = img[:,::-1,:] |
| batch_data = process_inputs(processor=processor, pixel_values=img, device='cuda') |
| |
| |
| |
| |
| replay_images.append(img) |
|
|
| |
| batch_data['pixel_values'] = batch_data['pixel_values'].unsqueeze(1) |
| |
| with torch.autocast("cuda", dtype=torch.bfloat16): |
| slot_outputs = get_current_slots(model, batch_data, texts, in_past_tokens=None, device_id='cuda') |
| |
| |
| with torch.autocast("cuda", dtype=torch.bfloat16): |
| actions = get_normalized_actions(model, batch_data, slot_outputs, device_id='cuda') |
| actions = actions.detach().float().cpu().numpy() |
| unnormalized_actions = denormalize_actions(model, cfg.unnorm_key, actions) |
| action = unnormalized_actions[0] |
|
|
| |
| action = normalize_gripper_action(action, binarize=True) |
|
|
| |
| |
| |
| |
|
|
| |
| obs, reward, done, info = env.step(action.tolist()) |
| t += 1 |
| if done: |
| task_successes += 1 |
| total_successes += 1 |
| break |
|
|
| task_episodes += 1 |
| total_episodes += 1 |
|
|
| |
| cv2.destroyAllWindows() |
| save_rollout_video( |
| replay_images, total_episodes, success=done, task_description=task_description, log_file=log_file, saved_dir=cfg.saved_dir |
| ) |
| |
| print(f"Success: {done}") |
| print(f"# episodes completed so far: {total_episodes}") |
| print(f"# successes: {total_successes} ({total_successes / total_episodes * 100:.1f}%)") |
| log_file.write(f"Success: {done}\n") |
| log_file.write(f"# episodes completed so far: {total_episodes}\n") |
| log_file.write(f"# successes: {total_successes} ({total_successes / total_episodes * 100:.1f}%)\n") |
| log_file.flush() |
| |
| |
| |
| |
| print(f"Current task success rate: {float(task_successes) / float(task_episodes)}") |
| print(f"Current total success rate: {float(total_successes) / float(total_episodes)}") |
| log_file.write(f"Current task success rate: {float(task_successes) / float(task_episodes)}\n") |
| log_file.write(f"Current total success rate: {float(total_successes) / float(total_episodes)}\n") |
| log_file.flush() |
|
|
| |
| log_file.close() |
|
|
| if __name__ == "__main__": |
| eval_libero() |
|
|