KV 融合实现完成。改动总结： 速度微弱提升psnr略微上升

attention.py — 3处改动：
  1. __init__ 添加 _kv_fused = False 标志
  2.新增 fuse_kv() 方法：将 to_k + to_v → to_kv，同时处理 _ip/_as/_aa 辅助 KV 对
  2. bmm_forward 两个分支加_kv_fused 判断，用to_kv().chunk(2, dim=-1) 替代分别调用

.claude/settings.local.json

@@ -0,0 +1,15 @@
+{
+  "permissions": {
+    "allow": [
+      "Bash(conda env list:*)",
+      "Bash(mamba env:*)",
+      "Bash(micromamba env list:*)",
+      "Bash(echo:*)",
+      "Bash(git show:*)",
+      "Bash(nvidia-smi:*)",
+      "Bash(conda activate unifolm-wma)",
+      "Bash(conda info:*)",
+      "Bash(direnv allow:*)"
+    ]
+  }
+}

.envrc

@@ -0,0 +1,2 @@
+eval "$(conda shell.bash hook 2>/dev/null)"
+conda activate unifolm-wma

.gitignore

@@ -55,7 +55,6 @@ coverage.xml
 *.pot
 
 # Django stuff:
-*.log
 local_settings.py
 db.sqlite3
 
@@ -121,7 +120,7 @@ localTest/
 fig/
 figure/
 *.mp4
-*.json
+
 Data/ControlVAE.yml
 Data/Misc
 Data/Pretrained
@@ -129,4 +128,7 @@ Data/utils.py
 Experiment/checkpoint
 Experiment/log
 
-*.ckpt
+*.ckpt
+
+*.0
+ckpts/unifolm_wma_dual.ckpt.prepared.pt

configs/inference/world_model_interaction.yaml

@@ -222,7 +222,7 @@ data:
     test:
       target: unifolm_wma.data.wma_data.WMAData
       params:
-        data_dir: '/mnt/ASC1637/unifolm-world-model-action/examples/world_model_interaction_prompts'
+        data_dir: '/home/qhy/unifolm-world-model-action/examples/world_model_interaction_prompts'
         video_length: ${model.params.wma_config.params.temporal_length}
         frame_stride: 2
         load_raw_resolution: True

scripts/evaluation/base_model_inference.py

@@ -16,6 +16,9 @@ from collections import OrderedDict
 from unifolm_wma.models.samplers.ddim import DDIMSampler
 from unifolm_wma.utils.utils import instantiate_from_config
 
+torch.backends.cuda.matmul.allow_tf32 = True
+torch.backends.cudnn.allow_tf32 = True
+
 
 def get_filelist(data_dir: str, postfixes: list[str]) -> list[str]:
     """

scripts/evaluation/real_eval_server.py

@@ -19,6 +19,9 @@ from fastapi.responses import JSONResponse
 from typing import Any, Dict, Optional, Tuple, List
 from datetime import datetime
 
+torch.backends.cuda.matmul.allow_tf32 = True
+torch.backends.cudnn.allow_tf32 = True
+
 from unifolm_wma.utils.utils import instantiate_from_config
 from unifolm_wma.models.samplers.ddim import DDIMSampler
 

scripts/evaluation/world_model_interaction.py

@@ -9,6 +9,8 @@ import logging
 import einops
 import warnings
 import imageio
+import atexit
+from concurrent.futures import ThreadPoolExecutor
 
 from pytorch_lightning import seed_everything
 from omegaconf import OmegaConf
@@ -16,8 +18,12 @@ from tqdm import tqdm
 from einops import rearrange, repeat
 from collections import OrderedDict
 from torch import nn
-from eval_utils import populate_queues, log_to_tensorboard
+from eval_utils import populate_queues
 from collections import deque
+from typing import Optional, List, Any
+
+torch.backends.cuda.matmul.allow_tf32 = True
+torch.backends.cudnn.allow_tf32 = True
 from torch import Tensor
 from torch.utils.tensorboard import SummaryWriter
 from PIL import Image
@@ -150,6 +156,81 @@ def save_results(video: Tensor, filename: str, fps: int = 8) -> None:
                                options={'crf': '10'})
 
 
+# ========== Async I/O ==========
+_io_executor: Optional[ThreadPoolExecutor] = None
+_io_futures: List[Any] = []
+
+
+def _get_io_executor() -> ThreadPoolExecutor:
+    global _io_executor
+    if _io_executor is None:
+        _io_executor = ThreadPoolExecutor(max_workers=2)
+    return _io_executor
+
+
+def _flush_io():
+    """Wait for all pending async I/O to finish."""
+    global _io_futures
+    for fut in _io_futures:
+        try:
+            fut.result()
+        except Exception as e:
+            print(f">>> [async I/O] error: {e}")
+    _io_futures.clear()
+
+
+atexit.register(_flush_io)
+
+
+def _save_results_sync(video_cpu: Tensor, filename: str, fps: int) -> None:
+    """Synchronous save on CPU tensor (runs in background thread)."""
+    video = torch.clamp(video_cpu.float(), -1., 1.)
+    n = video.shape[0]
+    video = video.permute(2, 0, 1, 3, 4)
+    frame_grids = [
+        torchvision.utils.make_grid(framesheet, nrow=int(n), padding=0)
+        for framesheet in video
+    ]
+    grid = torch.stack(frame_grids, dim=0)
+    grid = (grid + 1.0) / 2.0
+    grid = (grid * 255).to(torch.uint8).permute(0, 2, 3, 1)
+    torchvision.io.write_video(filename,
+                               grid,
+                               fps=fps,
+                               video_codec='h264',
+                               options={'crf': '10'})
+
+
+def save_results_async(video: Tensor, filename: str, fps: int = 8) -> None:
+    """Submit video saving to background thread pool."""
+    video_cpu = video.detach().cpu()
+    fut = _get_io_executor().submit(_save_results_sync, video_cpu, filename, fps)
+    _io_futures.append(fut)
+
+
+def _log_to_tb_sync(writer, video_cpu: Tensor, tag: str, fps: int) -> None:
+    """Synchronous TensorBoard log on CPU tensor (runs in background thread)."""
+    if video_cpu.dim() == 5:
+        n = video_cpu.shape[0]
+        video = video_cpu.permute(2, 0, 1, 3, 4)
+        frame_grids = [
+            torchvision.utils.make_grid(framesheet, nrow=int(n), padding=0)
+            for framesheet in video
+        ]
+        grid = torch.stack(frame_grids, dim=0)
+        grid = (grid + 1.0) / 2.0
+        grid = grid.unsqueeze(dim=0)
+        writer.add_video(tag, grid, fps=fps)
+
+
+def log_to_tensorboard_async(writer, data: Tensor, tag: str, fps: int = 10) -> None:
+    """Submit TensorBoard logging to background thread pool."""
+    if isinstance(data, torch.Tensor) and data.dim() == 5:
+        data_cpu = data.detach().cpu()
+        fut = _get_io_executor().submit(_log_to_tb_sync, writer, data_cpu, tag, fps)
+        _io_futures.append(fut)
+
+
 def get_init_frame_path(data_dir: str, sample: dict) -> str:
     """Construct the init_frame path from directory and sample metadata.
 
@@ -327,7 +408,8 @@ def image_guided_synthesis_sim_mode(
         timestep_spacing: str = 'uniform',
         guidance_rescale: float = 0.0,
         sim_mode: bool = True,
-        **kwargs) -> tuple[torch.Tensor, torch.Tensor, torch.Tensor]:
+        decode_video: bool = True,
+        **kwargs) -> tuple[torch.Tensor | None, torch.Tensor, torch.Tensor]:
     """
     Performs image-guided video generation in a simulation-style mode with optional multimodal guidance (image, state, action, text).
 
@@ -350,10 +432,13 @@ def image_guided_synthesis_sim_mode(
         timestep_spacing (str): Timestep sampling method in DDIM sampler. Typically "uniform" or "linspace".
         guidance_rescale (float): Guidance rescaling factor to mitigate overexposure from classifier-free guidance.
         sim_mode (bool): Whether to perform world-model interaction or decision-making using the world-model.
+        decode_video (bool): Whether to decode latent samples to pixel-space video.
+            Set to False to skip VAE decode for speed when only actions/states are needed.
         **kwargs: Additional arguments passed to the DDIM sampler.
 
     Returns:
-        batch_variants (torch.Tensor): Predicted pixel-space video frames [B, C, T, H, W].
+        batch_variants (torch.Tensor | None): Predicted pixel-space video frames [B, C, T, H, W],
+            or None when decode_video=False.
         actions (torch.Tensor): Predicted action sequences [B, T, D] from diffusion decoding.
         states (torch.Tensor): Predicted state sequences [B, T, D] from diffusion decoding.
     """
@@ -406,6 +491,7 @@ def image_guided_synthesis_sim_mode(
     kwargs.update({"unconditional_conditioning_img_nonetext": None})
     cond_mask = None
     cond_z0 = None
+    batch_variants = None
     if ddim_sampler is not None:
         samples, actions, states, intermedia = ddim_sampler.sample(
             S=ddim_steps,
@@ -424,9 +510,10 @@ def image_guided_synthesis_sim_mode(
             guidance_rescale=guidance_rescale,
             **kwargs)
 
-        # Reconstruct from latent to pixel space
-        batch_images = model.decode_first_stage(samples)
-        batch_variants = batch_images
+        if decode_video:
+            # Reconstruct from latent to pixel space
+            batch_images = model.decode_first_stage(samples)
+            batch_variants = batch_images
 
     return batch_variants, actions, states
 
@@ -453,26 +540,51 @@ def run_inference(args: argparse.Namespace, gpu_num: int, gpu_no: int) -> None:
     csv_path = os.path.join(args.prompt_dir, f"{args.dataset}.csv")
     df = pd.read_csv(csv_path)
 
-    # Load config
+    # Load config (always needed for data setup)
     config = OmegaConf.load(args.config)
-    config['model']['params']['wma_config']['params'][
-        'use_checkpoint'] = False
-    model = instantiate_from_config(config.model)
-    model.perframe_ae = args.perframe_ae
-    assert os.path.exists(args.ckpt_path), "Error: checkpoint Not Found!"
-    model = load_model_checkpoint(model, args.ckpt_path)
-    model.eval()
-    print(f'>>> Load pre-trained model ...')
 
-    # Build unnomalizer
+    prepared_path = args.ckpt_path + ".prepared.pt"
+    if os.path.exists(prepared_path):
+        # ---- Fast path: load the fully-prepared model ----
+        print(f">>> Loading prepared model from {prepared_path} ...")
+        model = torch.load(prepared_path,
+                           map_location=f"cuda:{gpu_no}",
+                           weights_only=False,
+                           mmap=True)
+        model.eval()
+        print(f">>> Prepared model loaded.")
+    else:
+        # ---- Normal path: construct + load checkpoint ----
+        config['model']['params']['wma_config']['params'][
+            'use_checkpoint'] = False
+        model = instantiate_from_config(config.model)
+        model.perframe_ae = args.perframe_ae
+
+        assert os.path.exists(args.ckpt_path), "Error: checkpoint Not Found!"
+        model = load_model_checkpoint(model, args.ckpt_path)
+        model.eval()
+        model = model.cuda(gpu_no)
+        print(f'>>> Load pre-trained model ...')
+
+        # Save prepared model for fast loading next time
+        print(f">>> Saving prepared model to {prepared_path} ...")
+        torch.save(model, prepared_path)
+        print(f">>> Prepared model saved ({os.path.getsize(prepared_path) / 1024**3:.1f} GB).")
+
+    # Build normalizer (always needed, independent of model loading path)
     logging.info("***** Configing Data *****")
     data = instantiate_from_config(config.data)
     data.setup()
     print(">>> Dataset is successfully loaded ...")
 
-    model = model.cuda(gpu_no)
     device = get_device_from_parameters(model)
 
+    # Fuse KV projections in attention layers (to_k + to_v → to_kv)
+    from unifolm_wma.modules.attention import CrossAttention
+    kv_count = sum(1 for m in model.modules()
+                   if isinstance(m, CrossAttention) and m.fuse_kv())
+    print(f"    ✓ KV fused: {kv_count} attention layers")
+
     # Run over data
     assert (args.height % 16 == 0) and (
         args.width % 16
@@ -587,7 +699,8 @@ def run_inference(args: argparse.Namespace, gpu_num: int, gpu_no: int) -> None:
                     fs=model_input_fs,
                     timestep_spacing=args.timestep_spacing,
                     guidance_rescale=args.guidance_rescale,
-                    sim_mode=False)
+                    sim_mode=False,
+                    decode_video=not args.fast_policy_no_decode)
 
                 # Update future actions in the observation queues
                 for idx in range(len(pred_actions[0])):
@@ -644,29 +757,31 @@ def run_inference(args: argparse.Namespace, gpu_num: int, gpu_no: int) -> None:
                     cond_obs_queues = populate_queues(cond_obs_queues,
                                                       observation)
 
-                # Save the imagen videos for decision-making
-                sample_tag = f"{args.dataset}-vid{sample['videoid']}-dm-fs-{fs}/itr-{itr}"
-                log_to_tensorboard(writer,
-                                   pred_videos_0,
-                                   sample_tag,
-                                   fps=args.save_fps)
+                # Save the imagen videos for decision-making (async)
+                if pred_videos_0 is not None:
+                    sample_tag = f"{args.dataset}-vid{sample['videoid']}-dm-fs-{fs}/itr-{itr}"
+                    log_to_tensorboard_async(writer,
+                                             pred_videos_0,
+                                             sample_tag,
+                                             fps=args.save_fps)
                 # Save videos environment changes via world-model interaction
                 sample_tag = f"{args.dataset}-vid{sample['videoid']}-wd-fs-{fs}/itr-{itr}"
-                log_to_tensorboard(writer,
-                                   pred_videos_1,
-                                   sample_tag,
-                                   fps=args.save_fps)
+                log_to_tensorboard_async(writer,
+                                         pred_videos_1,
+                                         sample_tag,
+                                         fps=args.save_fps)
 
                 # Save the imagen videos for decision-making
-                sample_video_file = f'{video_save_dir}/dm/{fs}/itr-{itr}.mp4'
-                save_results(pred_videos_0.cpu(),
-                             sample_video_file,
-                             fps=args.save_fps)
+                if pred_videos_0 is not None:
+                    sample_video_file = f'{video_save_dir}/dm/{fs}/itr-{itr}.mp4'
+                    save_results_async(pred_videos_0,
+                                       sample_video_file,
+                                       fps=args.save_fps)
                 # Save videos environment changes via world-model interaction
                 sample_video_file = f'{video_save_dir}/wm/{fs}/itr-{itr}.mp4'
-                save_results(pred_videos_1.cpu(),
-                             sample_video_file,
-                             fps=args.save_fps)
+                save_results_async(pred_videos_1,
+                                   sample_video_file,
+                                   fps=args.save_fps)
 
                 print('>' * 24)
                 # Collect the result of world-model interactions
@@ -674,12 +789,15 @@ def run_inference(args: argparse.Namespace, gpu_num: int, gpu_no: int) -> None:
 
             full_video = torch.cat(wm_video, dim=2)
             sample_tag = f"{args.dataset}-vid{sample['videoid']}-wd-fs-{fs}/full"
-            log_to_tensorboard(writer,
-                               full_video,
-                               sample_tag,
-                               fps=args.save_fps)
+            log_to_tensorboard_async(writer,
+                                     full_video,
+                                     sample_tag,
+                                     fps=args.save_fps)
             sample_full_video_file = f"{video_save_dir}/../{sample['videoid']}_full_fs{fs}.mp4"
-            save_results(full_video, sample_full_video_file, fps=args.save_fps)
+            save_results_async(full_video, sample_full_video_file, fps=args.save_fps)
+
+    # Wait for all async I/O to complete
+    _flush_io()
 
 
 def get_parser():
@@ -794,6 +912,11 @@ def get_parser():
                         action='store_true',
                         default=False,
                         help="not using the predicted states as comparison")
+    parser.add_argument(
+        "--fast_policy_no_decode",
+        action='store_true',
+        default=False,
+        help="Speed mode: policy pass only predicts actions, skip policy video decode/log/save.")
     parser.add_argument("--save_fps",
                         type=int,
                         default=8,

scripts/trainer.py

@@ -11,6 +11,9 @@ from unifolm_wma.utils.utils import instantiate_from_config
 from unifolm_wma.utils.train import get_trainer_callbacks, get_trainer_logger, get_trainer_strategy
 from unifolm_wma.utils.train import set_logger, init_workspace, load_checkpoints, get_num_parameters
 
+torch.backends.cuda.matmul.allow_tf32 = True
+torch.backends.cudnn.allow_tf32 = True
+
 
 def get_parser(**parser_kwargs):
     parser = argparse.ArgumentParser(**parser_kwargs)

src/unifolm_wma/models/diffusion_head/conditional_unet1d.py

@@ -501,6 +501,10 @@ class ConditionalUnet1D(nn.Module):
         self.last_frame_only = last_frame_only
         self.horizon = horizon
 
+        # Context precomputation cache
+        self._global_cond_cache_enabled = False
+        self._global_cond_cache = {}
+
     def forward(self,
                 sample: torch.Tensor,
                 timestep: Union[torch.Tensor, float, int],
@@ -530,14 +534,20 @@ class ConditionalUnet1D(nn.Module):
         B, T, D = sample.shape
         if self.use_linear_act_proj:
             sample = self.proj_in_action(sample.unsqueeze(-1))
-            global_cond = self.obs_encoder(cond)
-            global_cond = rearrange(global_cond,
-                                    '(b t) d -> b 1 (t d)',
-                                    b=B,
-                                    t=self.n_obs_steps)
-            global_cond = repeat(global_cond,
-                                 'b c d -> b (repeat c) d',
-                                 repeat=T)
+            _gc_key = (cond['image'].data_ptr(), cond['agent_pos'].data_ptr())
+            if self._global_cond_cache_enabled and _gc_key in self._global_cond_cache:
+                global_cond = self._global_cond_cache[_gc_key]
+            else:
+                global_cond = self.obs_encoder(cond)
+                global_cond = rearrange(global_cond,
+                                        '(b t) d -> b 1 (t d)',
+                                        b=B,
+                                        t=self.n_obs_steps)
+                global_cond = repeat(global_cond,
+                                     'b c d -> b (repeat c) d',
+                                     repeat=T)
+                if self._global_cond_cache_enabled:
+                    self._global_cond_cache[_gc_key] = global_cond
         else:
             sample = einops.rearrange(sample, 'b h t -> b t h')
             sample = self.proj_in_horizon(sample)

src/unifolm_wma/models/samplers/ddim.py

@@ -6,6 +6,8 @@ from unifolm_wma.utils.diffusion import make_ddim_sampling_parameters, make_ddim
 from unifolm_wma.utils.common import noise_like
 from unifolm_wma.utils.common import extract_into_tensor
 from tqdm import tqdm
+from unifolm_wma.modules.attention import enable_cross_attn_kv_cache, disable_cross_attn_kv_cache
+from unifolm_wma.modules.networks.wma_model import enable_ctx_cache, disable_ctx_cache
 
 
 class DDIMSampler(object):
@@ -67,11 +69,12 @@ class DDIMSampler(object):
             ddim_timesteps=self.ddim_timesteps,
             eta=ddim_eta,
             verbose=verbose)
-        self.register_buffer('ddim_sigmas', ddim_sigmas)
-        self.register_buffer('ddim_alphas', ddim_alphas)
-        self.register_buffer('ddim_alphas_prev', ddim_alphas_prev)
+        # Ensure tensors are on correct device for efficient indexing
+        self.register_buffer('ddim_sigmas', to_torch(torch.as_tensor(ddim_sigmas)))
+        self.register_buffer('ddim_alphas', to_torch(torch.as_tensor(ddim_alphas)))
+        self.register_buffer('ddim_alphas_prev', to_torch(torch.as_tensor(ddim_alphas_prev)))
         self.register_buffer('ddim_sqrt_one_minus_alphas',
-                             np.sqrt(1. - ddim_alphas))
+                             to_torch(torch.as_tensor(np.sqrt(1. - ddim_alphas))))
         sigmas_for_original_sampling_steps = ddim_eta * torch.sqrt(
             (1 - self.alphas_cumprod_prev) / (1 - self.alphas_cumprod) *
             (1 - self.alphas_cumprod / self.alphas_cumprod_prev))
@@ -241,63 +244,70 @@ class DDIMSampler(object):
 
         dp_ddim_scheduler_action.set_timesteps(len(timesteps))
         dp_ddim_scheduler_state.set_timesteps(len(timesteps))
-        for i, step in enumerate(iterator):
-            index = total_steps - i - 1
-            ts = torch.full((b, ), step, device=device, dtype=torch.long)
+        ts = torch.empty((b, ), device=device, dtype=torch.long)
+        enable_cross_attn_kv_cache(self.model)
+        enable_ctx_cache(self.model)
+        try:
+            for i, step in enumerate(iterator):
+                index = total_steps - i - 1
+                ts.fill_(step)
 
-            # Use mask to blend noised original latent (img_orig) & new sampled latent (img)
-            if mask is not None:
-                assert x0 is not None
-                if clean_cond:
-                    img_orig = x0
-                else:
-                    img_orig = self.model.q_sample(x0, ts)
-                img = img_orig * mask + (1. - mask) * img
+                # Use mask to blend noised original latent (img_orig) & new sampled latent (img)
+                if mask is not None:
+                    assert x0 is not None
+                    if clean_cond:
+                        img_orig = x0
+                    else:
+                        img_orig = self.model.q_sample(x0, ts)
+                    img = img_orig * mask + (1. - mask) * img
 
-            outs = self.p_sample_ddim(
-                img,
-                action,
-                state,
-                cond,
-                ts,
-                index=index,
-                use_original_steps=ddim_use_original_steps,
-                quantize_denoised=quantize_denoised,
-                temperature=temperature,
-                noise_dropout=noise_dropout,
-                score_corrector=score_corrector,
-                corrector_kwargs=corrector_kwargs,
-                unconditional_guidance_scale=unconditional_guidance_scale,
-                unconditional_conditioning=unconditional_conditioning,
-                mask=mask,
-                x0=x0,
-                fs=fs,
-                guidance_rescale=guidance_rescale,
-                **kwargs)
+                outs = self.p_sample_ddim(
+                    img,
+                    action,
+                    state,
+                    cond,
+                    ts,
+                    index=index,
+                    use_original_steps=ddim_use_original_steps,
+                    quantize_denoised=quantize_denoised,
+                    temperature=temperature,
+                    noise_dropout=noise_dropout,
+                    score_corrector=score_corrector,
+                    corrector_kwargs=corrector_kwargs,
+                    unconditional_guidance_scale=unconditional_guidance_scale,
+                    unconditional_conditioning=unconditional_conditioning,
+                    mask=mask,
+                    x0=x0,
+                    fs=fs,
+                    guidance_rescale=guidance_rescale,
+                    **kwargs)
 
-            img, pred_x0, model_output_action, model_output_state = outs
+                img, pred_x0, model_output_action, model_output_state = outs
 
-            action = dp_ddim_scheduler_action.step(
-                model_output_action,
-                step,
-                action,
-                generator=None,
-            ).prev_sample
-            state = dp_ddim_scheduler_state.step(
-                model_output_state,
-                step,
-                state,
-                generator=None,
-            ).prev_sample
+                action = dp_ddim_scheduler_action.step(
+                    model_output_action,
+                    step,
+                    action,
+                    generator=None,
+                ).prev_sample
+                state = dp_ddim_scheduler_state.step(
+                    model_output_state,
+                    step,
+                    state,
+                    generator=None,
+                ).prev_sample
 
-            if callback: callback(i)
-            if img_callback: img_callback(pred_x0, i)
+                if callback: callback(i)
+                if img_callback: img_callback(pred_x0, i)
 
-            if index % log_every_t == 0 or index == total_steps - 1:
-                intermediates['x_inter'].append(img)
-                intermediates['pred_x0'].append(pred_x0)
-                intermediates['x_inter_action'].append(action)
-                intermediates['x_inter_state'].append(state)
+                if index % log_every_t == 0 or index == total_steps - 1:
+                    intermediates['x_inter'].append(img)
+                    intermediates['pred_x0'].append(pred_x0)
+                    intermediates['x_inter_action'].append(action)
+                    intermediates['x_inter_state'].append(state)
+        finally:
+            disable_cross_attn_kv_cache(self.model)
+            disable_ctx_cache(self.model)
 
         return img, action, state, intermediates
 
@@ -325,10 +335,6 @@ class DDIMSampler(object):
                       guidance_rescale=0.0,
                       **kwargs):
         b, *_, device = *x.shape, x.device
-        if x.dim() == 5:
-            is_video = True
-        else:
-            is_video = False
 
         if unconditional_conditioning is None or unconditional_guidance_scale == 1.:
             model_output, model_output_action, model_output_state = self.model.apply_model(
@@ -377,17 +383,11 @@ class DDIMSampler(object):
         sqrt_one_minus_alphas = self.model.sqrt_one_minus_alphas_cumprod if use_original_steps else self.ddim_sqrt_one_minus_alphas
         sigmas = self.ddim_sigmas_for_original_num_steps if use_original_steps else self.ddim_sigmas
 
-        if is_video:
-            size = (b, 1, 1, 1, 1)
-        else:
-            size = (b, 1, 1, 1)
-
-        a_t = torch.full(size, alphas[index], device=device)
-        a_prev = torch.full(size, alphas_prev[index], device=device)
-        sigma_t = torch.full(size, sigmas[index], device=device)
-        sqrt_one_minus_at = torch.full(size,
-                                       sqrt_one_minus_alphas[index],
-                                       device=device)
+        # Use 0-d tensors directly (already on device); broadcasting handles shape
+        a_t = alphas[index]
+        a_prev = alphas_prev[index]
+        sigma_t = sigmas[index]
+        sqrt_one_minus_at = sqrt_one_minus_alphas[index]
 
         if self.model.parameterization != "v":
             pred_x0 = (x - sqrt_one_minus_at * e_t) / a_t.sqrt()
@@ -395,12 +395,8 @@ class DDIMSampler(object):
             pred_x0 = self.model.predict_start_from_z_and_v(x, t, model_output)
 
         if self.model.use_dynamic_rescale:
-            scale_t = torch.full(size,
-                                 self.ddim_scale_arr[index],
-                                 device=device)
-            prev_scale_t = torch.full(size,
-                                      self.ddim_scale_arr_prev[index],
-                                      device=device)
+            scale_t = self.ddim_scale_arr[index]
+            prev_scale_t = self.ddim_scale_arr_prev[index]
             rescale = (prev_scale_t / scale_t)
             pred_x0 *= rescale
 

src/unifolm_wma/modules/attention.py

@@ -98,6 +98,10 @@ class CrossAttention(nn.Module):
         self.text_context_len = text_context_len
         self.agent_state_context_len = agent_state_context_len
         self.agent_action_context_len = agent_action_context_len
+        self._kv_cache = {}
+        self._kv_cache_enabled = False
+        self._kv_fused = False
+
         self.cross_attention_scale_learnable = cross_attention_scale_learnable
         if self.image_cross_attention:
             self.to_k_ip = nn.Linear(context_dim, inner_dim, bias=False)
@@ -114,6 +118,27 @@ class CrossAttention(nn.Module):
                 self.register_parameter('alpha_caa',
                                         nn.Parameter(torch.tensor(0.)))
 
+    def fuse_kv(self):
+        """Fuse to_k/to_v into to_kv (2 Linear → 1). Works for all layers."""
+        k_w = self.to_k.weight  # (inner_dim, context_dim)
+        v_w = self.to_v.weight
+        self.to_kv = nn.Linear(k_w.shape[1], k_w.shape[0] * 2, bias=False)
+        self.to_kv.weight = nn.Parameter(torch.cat([k_w, v_w], dim=0))
+        del self.to_k, self.to_v
+        if self.image_cross_attention:
+            for suffix in ('_ip', '_as', '_aa'):
+                k_attr = f'to_k{suffix}'
+                v_attr = f'to_v{suffix}'
+                kw = getattr(self, k_attr).weight
+                vw = getattr(self, v_attr).weight
+                fused = nn.Linear(kw.shape[1], kw.shape[0] * 2, bias=False)
+                fused.weight = nn.Parameter(torch.cat([kw, vw], dim=0))
+                setattr(self, f'to_kv{suffix}', fused)
+                delattr(self, k_attr)
+                delattr(self, v_attr)
+        self._kv_fused = True
+        return True
+
     def forward(self, x, context=None, mask=None):
         spatial_self_attn = (context is None)
         k_ip, v_ip, out_ip = None, None, None
@@ -140,19 +165,28 @@ class CrossAttention(nn.Module):
                                     self.agent_action_context_len +
                                     self.text_context_len:, :]
 
-            k = self.to_k(context_ins)
-            v = self.to_v(context_ins)
-            k_ip = self.to_k_ip(context_image)
-            v_ip = self.to_v_ip(context_image)
-            k_as = self.to_k_as(context_agent_state)
-            v_as = self.to_v_as(context_agent_state)
-            k_aa = self.to_k_aa(context_agent_action)
-            v_aa = self.to_v_aa(context_agent_action)
+            if self._kv_fused:
+                k, v = self.to_kv(context_ins).chunk(2, dim=-1)
+                k_ip, v_ip = self.to_kv_ip(context_image).chunk(2, dim=-1)
+                k_as, v_as = self.to_kv_as(context_agent_state).chunk(2, dim=-1)
+                k_aa, v_aa = self.to_kv_aa(context_agent_action).chunk(2, dim=-1)
+            else:
+                k = self.to_k(context_ins)
+                v = self.to_v(context_ins)
+                k_ip = self.to_k_ip(context_image)
+                v_ip = self.to_v_ip(context_image)
+                k_as = self.to_k_as(context_agent_state)
+                v_as = self.to_v_as(context_agent_state)
+                k_aa = self.to_k_aa(context_agent_action)
+                v_aa = self.to_v_aa(context_agent_action)
         else:
             if not spatial_self_attn:
                 context = context[:, :self.text_context_len, :]
-            k = self.to_k(context)
-            v = self.to_v(context)
+            if self._kv_fused:
+                k, v = self.to_kv(context).chunk(2, dim=-1)
+            else:
+                k = self.to_k(context)
+                v = self.to_v(context)
 
         q, k, v = map(lambda t: rearrange(t, 'b n (h d) -> (b h) n d', h=h),
                       (q, k, v))
@@ -236,134 +270,162 @@ class CrossAttention(nn.Module):
         k_ip, v_ip, out_ip = None, None, None
         k_as, v_as, out_as = None, None, None
         k_aa, v_aa, out_aa = None, None, None
+        attn_mask_aa = None
 
+        h = self.heads
         q = self.to_q(x)
         context = default(context, x)
 
-        if self.image_cross_attention and not spatial_self_attn:
+        b, _, _ = q.shape
+        q = q.unsqueeze(3).reshape(b, q.shape[1], h, self.dim_head).permute(0, 2, 1, 3).reshape(b * h, q.shape[1], self.dim_head).contiguous()
+
+        def _reshape_kv(t):
+            return t.unsqueeze(3).reshape(b, t.shape[1], h, self.dim_head).permute(0, 2, 1, 3).reshape(b * h, t.shape[1], self.dim_head).contiguous()
+
+        use_cache = self._kv_cache_enabled and not spatial_self_attn
+        cache_hit = use_cache and len(self._kv_cache) > 0
+
+        if cache_hit:
+            k = self._kv_cache['k']
+            v = self._kv_cache['v']
+            k_ip = self._kv_cache.get('k_ip')
+            v_ip = self._kv_cache.get('v_ip')
+            k_as = self._kv_cache.get('k_as')
+            v_as = self._kv_cache.get('v_as')
+            k_aa = self._kv_cache.get('k_aa')
+            v_aa = self._kv_cache.get('v_aa')
+            attn_mask_aa = self._kv_cache.get('attn_mask_aa')
+        elif self.image_cross_attention and not spatial_self_attn:
             if context.shape[1] == self.text_context_len + self.video_length:
                 context_ins, context_image = context[:, :self.text_context_len, :], context[:,self.text_context_len:, :]
-                k = self.to_k(context)
-                v = self.to_v(context)
-                k_ip = self.to_k_ip(context_image)
-                v_ip = self.to_v_ip(context_image)
+                if self._kv_fused:
+                    k, v = self.to_kv(context).chunk(2, dim=-1)
+                    k_ip, v_ip = self.to_kv_ip(context_image).chunk(2, dim=-1)
+                else:
+                    k = self.to_k(context)
+                    v = self.to_v(context)
+                    k_ip = self.to_k_ip(context_image)
+                    v_ip = self.to_v_ip(context_image)
+                k, v = map(_reshape_kv, (k, v))
+                k_ip, v_ip = map(_reshape_kv, (k_ip, v_ip))
+                if use_cache:
+                    self._kv_cache = {'k': k, 'v': v, 'k_ip': k_ip, 'v_ip': v_ip}
             elif context.shape[1] == self.agent_state_context_len + self.text_context_len + self.video_length:
                 context_agent_state = context[:, :self.agent_state_context_len, :]
                 context_ins = context[:, self.agent_state_context_len:self.agent_state_context_len+self.text_context_len, :]
                 context_image = context[:, self.agent_state_context_len+self.text_context_len:, :]
-                k = self.to_k(context_ins)
-                v = self.to_v(context_ins)
-                k_ip = self.to_k_ip(context_image)
-                v_ip = self.to_v_ip(context_image)
-                k_as = self.to_k_as(context_agent_state)
-                v_as = self.to_v_as(context_agent_state)
+                if self._kv_fused:
+                    k, v = self.to_kv(context_ins).chunk(2, dim=-1)
+                    k_ip, v_ip = self.to_kv_ip(context_image).chunk(2, dim=-1)
+                    k_as, v_as = self.to_kv_as(context_agent_state).chunk(2, dim=-1)
+                else:
+                    k = self.to_k(context_ins)
+                    v = self.to_v(context_ins)
+                    k_ip = self.to_k_ip(context_image)
+                    v_ip = self.to_v_ip(context_image)
+                    k_as = self.to_k_as(context_agent_state)
+                    v_as = self.to_v_as(context_agent_state)
+                k, v = map(_reshape_kv, (k, v))
+                k_ip, v_ip = map(_reshape_kv, (k_ip, v_ip))
+                k_as, v_as = map(_reshape_kv, (k_as, v_as))
+                if use_cache:
+                    self._kv_cache = {'k': k, 'v': v, 'k_ip': k_ip, 'v_ip': v_ip, 'k_as': k_as, 'v_as': v_as}
             else:
                 context_agent_state = context[:, :self.agent_state_context_len, :]
                 context_agent_action = context[:, self.agent_state_context_len:self.agent_state_context_len+self.agent_action_context_len, :]
                 context_ins = context[:, self.agent_state_context_len+self.agent_action_context_len:self.agent_state_context_len+self.agent_action_context_len+self.text_context_len, :]
                 context_image = context[:, self.agent_state_context_len+self.agent_action_context_len+self.text_context_len:, :]
 
-                k = self.to_k(context_ins)
-                v = self.to_v(context_ins)
-                k_ip = self.to_k_ip(context_image)
-                v_ip = self.to_v_ip(context_image)
-                k_as = self.to_k_as(context_agent_state)
-                v_as = self.to_v_as(context_agent_state)
-                k_aa = self.to_k_aa(context_agent_action)
-                v_aa = self.to_v_aa(context_agent_action)
+                if self._kv_fused:
+                    k, v = self.to_kv(context_ins).chunk(2, dim=-1)
+                    k_ip, v_ip = self.to_kv_ip(context_image).chunk(2, dim=-1)
+                    k_as, v_as = self.to_kv_as(context_agent_state).chunk(2, dim=-1)
+                    k_aa, v_aa = self.to_kv_aa(context_agent_action).chunk(2, dim=-1)
+                else:
+                    k = self.to_k(context_ins)
+                    v = self.to_v(context_ins)
+                    k_ip = self.to_k_ip(context_image)
+                    v_ip = self.to_v_ip(context_image)
+                    k_as = self.to_k_as(context_agent_state)
+                    v_as = self.to_v_as(context_agent_state)
+                    k_aa = self.to_k_aa(context_agent_action)
+                    v_aa = self.to_v_aa(context_agent_action)
 
-                attn_mask_aa = self._get_attn_mask_aa(x.shape[0],
-                                                      q.shape[1],
-                                                      k_aa.shape[1],
-                                                      block_size=16).to(k_aa.device)
+                k, v = map(_reshape_kv, (k, v))
+                k_ip, v_ip = map(_reshape_kv, (k_ip, v_ip))
+                k_as, v_as = map(_reshape_kv, (k_as, v_as))
+                k_aa, v_aa = map(_reshape_kv, (k_aa, v_aa))
+
+                attn_mask_aa_raw = self._get_attn_mask_aa(x.shape[0],
+                                                          q.shape[1],
+                                                          k_aa.shape[1],
+                                                          block_size=16,
+                                                          device=k_aa.device)
+                attn_mask_aa = attn_mask_aa_raw.unsqueeze(1).repeat(1, h, 1, 1).reshape(
+                    b * h, attn_mask_aa_raw.shape[1], attn_mask_aa_raw.shape[2]).to(q.dtype)
+
+                if use_cache:
+                    self._kv_cache = {
+                        'k': k, 'v': v, 'k_ip': k_ip, 'v_ip': v_ip,
+                        'k_as': k_as, 'v_as': v_as, 'k_aa': k_aa, 'v_aa': v_aa,
+                        'attn_mask_aa': attn_mask_aa,
+                    }
         else:
             if not spatial_self_attn:
                 assert 1 > 2, ">>> ERROR: you should never go into here ..."
                 context = context[:, :self.text_context_len, :]
-            k = self.to_k(context)
-            v = self.to_v(context)
-
-        b, _, _ = q.shape
-        q = q.unsqueeze(3).reshape(b, q.shape[1], self.heads, self.dim_head).permute(0, 2, 1, 3).reshape(b * self.heads, q.shape[1], self.dim_head).contiguous()
+            if self._kv_fused:
+                k, v = self.to_kv(context).chunk(2, dim=-1)
+            else:
+                k = self.to_k(context)
+                v = self.to_v(context)
+            k, v = map(_reshape_kv, (k, v))
+            if use_cache:
+                self._kv_cache = {'k': k, 'v': v}
         if k is not None:
-            k, v = map(
-                lambda t: t.unsqueeze(3).reshape(b, t.shape[
-                    1], self.heads, self.dim_head).permute(0, 2, 1, 3).reshape(
-                        b * self.heads, t.shape[1], self.dim_head).contiguous(),
-                (k, v),
-            )
             out = xformers.ops.memory_efficient_attention(q,
                                                           k,
                                                           v,
                                                           attn_bias=None,
                                                           op=None)
             out = (out.unsqueeze(0).reshape(
-                b, self.heads, out.shape[1],
+                b, h, out.shape[1],
                 self.dim_head).permute(0, 2, 1,
                                        3).reshape(b, out.shape[1],
-                                                  self.heads * self.dim_head))
+                                                  h * self.dim_head))
 
         if k_ip is not None:
-            # For image cross-attention
-            k_ip, v_ip = map(
-                lambda t: t.unsqueeze(3).reshape(b, t.shape[
-                    1], self.heads, self.dim_head).permute(0, 2, 1, 3).reshape(
-                        b * self.heads, t.shape[1], self.dim_head).contiguous(
-                        ),
-                (k_ip, v_ip),
-            )
             out_ip = xformers.ops.memory_efficient_attention(q,
                                                              k_ip,
                                                              v_ip,
                                                              attn_bias=None,
                                                              op=None)
             out_ip = (out_ip.unsqueeze(0).reshape(
-                b, self.heads, out_ip.shape[1],
+                b, h, out_ip.shape[1],
                 self.dim_head).permute(0, 2, 1,
                                        3).reshape(b, out_ip.shape[1],
-                                                  self.heads * self.dim_head))
+                                                  h * self.dim_head))
 
         if k_as is not None:
-            # For agent state cross-attention
-            k_as, v_as = map(
-                lambda t: t.unsqueeze(3).reshape(b, t.shape[
-                    1], self.heads, self.dim_head).permute(0, 2, 1, 3).reshape(
-                        b * self.heads, t.shape[1], self.dim_head).contiguous(
-                        ),
-                (k_as, v_as),
-            )
             out_as = xformers.ops.memory_efficient_attention(q,
                                                              k_as,
                                                              v_as,
                                                              attn_bias=None,
                                                              op=None)
             out_as = (out_as.unsqueeze(0).reshape(
-                b, self.heads, out_as.shape[1],
+                b, h, out_as.shape[1],
                 self.dim_head).permute(0, 2, 1,
                                        3).reshape(b, out_as.shape[1],
-                                                  self.heads * self.dim_head))
+                                                  h * self.dim_head))
+
         if k_aa is not None:
-            # For agent action cross-attention
-            k_aa, v_aa = map(
-                lambda t: t.unsqueeze(3).reshape(b, t.shape[
-                    1], self.heads, self.dim_head).permute(0, 2, 1, 3).reshape(
-                        b * self.heads, t.shape[1], self.dim_head).contiguous(
-                        ),
-                (k_aa, v_aa),
-            )
-
-            attn_mask_aa = attn_mask_aa.unsqueeze(1).repeat(1,self.heads,1,1).reshape(
-                    b * self.heads, attn_mask_aa.shape[1], attn_mask_aa.shape[2])
-            attn_mask_aa = attn_mask_aa.to(q.dtype)
-
             out_aa = xformers.ops.memory_efficient_attention(
                 q, k_aa, v_aa, attn_bias=attn_mask_aa, op=None)
-
             out_aa = (out_aa.unsqueeze(0).reshape(
-                b, self.heads, out_aa.shape[1],
+                b, h, out_aa.shape[1],
                 self.dim_head).permute(0, 2, 1,
                                        3).reshape(b, out_aa.shape[1],
-                                                  self.heads * self.dim_head))
+                                                  h * self.dim_head))
         if exists(mask):
             raise NotImplementedError
 
@@ -386,17 +448,43 @@ class CrossAttention(nn.Module):
 
         return self.to_out(out)
 
-    def _get_attn_mask_aa(self, b, l1, l2, block_size=16):
+    def _get_attn_mask_aa(self, b, l1, l2, block_size=16, device=None):
+        cache_key = (b, l1, l2, block_size)
+        if hasattr(self, '_attn_mask_aa_cache_key') and self._attn_mask_aa_cache_key == cache_key:
+            cached = self._attn_mask_aa_cache
+            if device is not None and cached.device != torch.device(device):
+                cached = cached.to(device)
+                self._attn_mask_aa_cache = cached
+            return cached
+
+        target_device = device if device is not None else 'cpu'
         num_token = l2 // block_size
-        start_positions = ((torch.arange(b) % block_size) + 1) * num_token
-        col_indices = torch.arange(l2)
+        start_positions = ((torch.arange(b, device=target_device) % block_size) + 1) * num_token
+        col_indices = torch.arange(l2, device=target_device)
         mask_2d = col_indices.unsqueeze(0) >= start_positions.unsqueeze(1)
         mask = mask_2d.unsqueeze(1).expand(b, l1, l2)
-        attn_mask = torch.zeros_like(mask, dtype=torch.float)
+        attn_mask = torch.zeros(b, l1, l2, dtype=torch.float, device=target_device)
         attn_mask[mask] = float('-inf')
+
+        self._attn_mask_aa_cache_key = cache_key
+        self._attn_mask_aa_cache = attn_mask
         return attn_mask
 
 
+def enable_cross_attn_kv_cache(module):
+    for m in module.modules():
+        if isinstance(m, CrossAttention):
+            m._kv_cache_enabled = True
+            m._kv_cache = {}
+
+
+def disable_cross_attn_kv_cache(module):
+    for m in module.modules():
+        if isinstance(m, CrossAttention):
+            m._kv_cache_enabled = False
+            m._kv_cache = {}
+
+
 class BasicTransformerBlock(nn.Module):
 
     def __init__(self,

src/unifolm_wma/modules/networks/wma_model.py

@@ -685,6 +685,21 @@ class WMAModel(nn.Module):
         self.action_token_projector = instantiate_from_config(
             stem_process_config)
 
+        # Context precomputation cache
+        self._ctx_cache_enabled = False
+        self._ctx_cache = {}
+        # Reusable CUDA stream for parallel state_unet / action_unet
+        self._state_stream = torch.cuda.Stream()
+
+    def __getstate__(self):
+        state = self.__dict__.copy()
+        state.pop('_state_stream', None)
+        return state
+
+    def __setstate__(self, state):
+        self.__dict__.update(state)
+        self._state_stream = torch.cuda.Stream()
+
     def forward(self,
                 x: Tensor,
                 x_action: Tensor,
@@ -720,58 +735,64 @@ class WMAModel(nn.Module):
                                    repeat_only=False).type(x.dtype)
         emb = self.time_embed(t_emb)
 
-        bt, l_context, _ = context.shape
-        if self.base_model_gen_only:
-            assert l_context == 77 + self.n_obs_steps * 16, ">>> ERROR Context dim 1 ..."  ## NOTE HANDCODE
+        _ctx_key = context.data_ptr()
+        if self._ctx_cache_enabled and _ctx_key in self._ctx_cache:
+            context = self._ctx_cache[_ctx_key]
         else:
-            if l_context == self.n_obs_steps + 77 + t * 16:
-                context_agent_state = context[:, :self.n_obs_steps]
-                context_text = context[:, self.n_obs_steps:self.n_obs_steps +
-                                       77, :]
-                context_img = context[:, self.n_obs_steps + 77:, :]
-                context_agent_state = context_agent_state.repeat_interleave(
-                    repeats=t, dim=0)
-                context_text = context_text.repeat_interleave(repeats=t, dim=0)
-                context_img = rearrange(context_img,
-                                        'b (t l) c -> (b t) l c',
-                                        t=t)
-                context = torch.cat(
-                    [context_agent_state, context_text, context_img], dim=1)
-            elif l_context == self.n_obs_steps + 16 + 77 + t * 16:
-                context_agent_state = context[:, :self.n_obs_steps]
-                context_agent_action = context[:, self.
-                                               n_obs_steps:self.n_obs_steps +
-                                               16, :]
-                context_agent_action = rearrange(
-                    context_agent_action.unsqueeze(2), 'b t l d -> (b t) l d')
-                context_agent_action = self.action_token_projector(
-                    context_agent_action)
-                context_agent_action = rearrange(context_agent_action,
-                                                 '(b o) l d -> b o l d',
-                                                 o=t)
-                context_agent_action = rearrange(context_agent_action,
-                                                 'b o (t l) d -> b o t l d',
-                                                 t=t)
-                context_agent_action = context_agent_action.permute(
-                    0, 2, 1, 3, 4)
-                context_agent_action = rearrange(context_agent_action,
-                                                 'b t o l d -> (b t) (o l) d')
+            bt, l_context, _ = context.shape
+            if self.base_model_gen_only:
+                assert l_context == 77 + self.n_obs_steps * 16, ">>> ERROR Context dim 1 ..."  ## NOTE HANDCODE
+            else:
+                if l_context == self.n_obs_steps + 77 + t * 16:
+                    context_agent_state = context[:, :self.n_obs_steps]
+                    context_text = context[:, self.n_obs_steps:self.n_obs_steps +
+                                           77, :]
+                    context_img = context[:, self.n_obs_steps + 77:, :]
+                    context_agent_state = context_agent_state.repeat_interleave(
+                        repeats=t, dim=0)
+                    context_text = context_text.repeat_interleave(repeats=t, dim=0)
+                    context_img = rearrange(context_img,
+                                            'b (t l) c -> (b t) l c',
+                                            t=t)
+                    context = torch.cat(
+                        [context_agent_state, context_text, context_img], dim=1)
+                elif l_context == self.n_obs_steps + 16 + 77 + t * 16:
+                    context_agent_state = context[:, :self.n_obs_steps]
+                    context_agent_action = context[:, self.
+                                                   n_obs_steps:self.n_obs_steps +
+                                                   16, :]
+                    context_agent_action = rearrange(
+                        context_agent_action.unsqueeze(2), 'b t l d -> (b t) l d')
+                    context_agent_action = self.action_token_projector(
+                        context_agent_action)
+                    context_agent_action = rearrange(context_agent_action,
+                                                     '(b o) l d -> b o l d',
+                                                     o=t)
+                    context_agent_action = rearrange(context_agent_action,
+                                                     'b o (t l) d -> b o t l d',
+                                                     t=t)
+                    context_agent_action = context_agent_action.permute(
+                        0, 2, 1, 3, 4)
+                    context_agent_action = rearrange(context_agent_action,
+                                                     'b t o l d -> (b t) (o l) d')
 
-                context_text = context[:, self.n_obs_steps +
-                                       16:self.n_obs_steps + 16 + 77, :]
-                context_text = context_text.repeat_interleave(repeats=t, dim=0)
+                    context_text = context[:, self.n_obs_steps +
+                                           16:self.n_obs_steps + 16 + 77, :]
+                    context_text = context_text.repeat_interleave(repeats=t, dim=0)
 
-                context_img = context[:, self.n_obs_steps + 16 + 77:, :]
-                context_img = rearrange(context_img,
-                                        'b (t l) c -> (b t) l c',
-                                        t=t)
-                context_agent_state = context_agent_state.repeat_interleave(
-                    repeats=t, dim=0)
-                context = torch.cat([
-                    context_agent_state, context_agent_action, context_text,
-                    context_img
-                ],
-                                    dim=1)
+                    context_img = context[:, self.n_obs_steps + 16 + 77:, :]
+                    context_img = rearrange(context_img,
+                                            'b (t l) c -> (b t) l c',
+                                            t=t)
+                    context_agent_state = context_agent_state.repeat_interleave(
+                        repeats=t, dim=0)
+                    context = torch.cat([
+                        context_agent_state, context_agent_action, context_text,
+                        context_img
+                    ],
+                                        dim=1)
+            if self._ctx_cache_enabled:
+                self._ctx_cache[_ctx_key] = context
 
         emb = emb.repeat_interleave(repeats=t, dim=0)
 
@@ -832,17 +853,45 @@ class WMAModel(nn.Module):
 
         if not self.base_model_gen_only:
             ba, _, _ = x_action.shape
+            ts_state = timesteps[:ba] if b > 1 else timesteps
+            # Run action_unet and state_unet in parallel via CUDA streams
+            s_stream = self._state_stream
+            s_stream.wait_stream(torch.cuda.current_stream())
+            with torch.cuda.stream(s_stream):
+                s_y = self.state_unet(x_state, ts_state, hs_a,
+                                      context_action[:2], **kwargs)
             a_y = self.action_unet(x_action, timesteps[:ba], hs_a,
                                    context_action[:2], **kwargs)
-            # Predict state
-            if b > 1:
-                s_y = self.state_unet(x_state, timesteps[:ba], hs_a,
-                                      context_action[:2], **kwargs)
-            else:
-                s_y = self.state_unet(x_state, timesteps, hs_a,
-                                      context_action[:2], **kwargs)
+            torch.cuda.current_stream().wait_stream(s_stream)
         else:
             a_y = torch.zeros_like(x_action)
             s_y = torch.zeros_like(x_state)
 
         return y, a_y, s_y
+
+
+def enable_ctx_cache(model):
+    """Enable context precomputation cache on WMAModel and its action/state UNets."""
+    for m in model.modules():
+        if isinstance(m, WMAModel):
+            m._ctx_cache_enabled = True
+            m._ctx_cache = {}
+    # conditional_unet1d cache
+    from unifolm_wma.models.diffusion_head.conditional_unet1d import ConditionalUnet1D
+    for m in model.modules():
+        if isinstance(m, ConditionalUnet1D):
+            m._global_cond_cache_enabled = True
+            m._global_cond_cache = {}
+
+
+def disable_ctx_cache(model):
+    """Disable and clear context precomputation cache."""
+    for m in model.modules():
+        if isinstance(m, WMAModel):
+            m._ctx_cache_enabled = False
+            m._ctx_cache = {}
+    from unifolm_wma.models.diffusion_head.conditional_unet1d import ConditionalUnet1D
+    for m in model.modules():
+        if isinstance(m, ConditionalUnet1D):
+            m._global_cond_cache_enabled = False
+            m._global_cond_cache = {}

unitree_z1_dual_arm_cleanup_pencils/case1/output.log

@@ -0,0 +1,121 @@
+2026-02-10 15:38:28.973314: I tensorflow/core/util/port.cc:113] oneDNN custom operations are on. You may see slightly different numerical results due to floating-point round-off errors from different computation orders. To turn them off, set the environment variable `TF_ENABLE_ONEDNN_OPTS=0`.
+2026-02-10 15:38:29.023024: E external/local_xla/xla/stream_executor/cuda/cuda_dnn.cc:9261] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered
+2026-02-10 15:38:29.023070: E external/local_xla/xla/stream_executor/cuda/cuda_fft.cc:607] Unable to register cuFFT factory: Attempting to register factory for plugin cuFFT when one has already been registered
+2026-02-10 15:38:29.024393: E external/local_xla/xla/stream_executor/cuda/cuda_blas.cc:1515] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered
+2026-02-10 15:38:29.031901: I tensorflow/core/platform/cpu_feature_guard.cc:182] This TensorFlow binary is optimized to use available CPU instructions in performance-critical operations.
+To enable the following instructions: AVX2 AVX512F AVX512_VNNI FMA, in other operations, rebuild TensorFlow with the appropriate compiler flags.
+2026-02-10 15:38:29.955454: W tensorflow/compiler/tf2tensorrt/utils/py_utils.cc:38] TF-TRT Warning: Could not find TensorRT
+Global seed set to 123
+INFO:mainlogger:LatentVisualDiffusion: Running in v-prediction mode
+INFO:unifolm_wma.models.diffusion_head.conditional_unet1d:number of parameters: 5.010531e+08
+INFO:unifolm_wma.models.diffusion_head.conditional_unet1d:number of parameters: 5.010531e+08
+AE working on z of shape (1, 4, 32, 32) = 4096 dimensions.
+INFO:root:Loaded ViT-H-14 model config.
+DEBUG:urllib3.connectionpool:Starting new HTTPS connection (1): hf-mirror.com:443
+DEBUG:urllib3.connectionpool:https://hf-mirror.com:443 "HEAD /laion/CLIP-ViT-H-14-laion2B-s32B-b79K/resolve/main/open_clip_pytorch_model.bin HTTP/1.1" 302 0
+INFO:root:Loading pretrained ViT-H-14 weights (laion2b_s32b_b79k).
+INFO:root:Loaded ViT-H-14 model config.
+DEBUG:urllib3.connectionpool:https://hf-mirror.com:443 "HEAD /laion/CLIP-ViT-H-14-laion2B-s32B-b79K/resolve/main/open_clip_pytorch_model.bin HTTP/1.1" 302 0
+INFO:root:Loading pretrained ViT-H-14 weights (laion2b_s32b_b79k).
+>>> model checkpoint loaded.
+>>> Load pre-trained model ...
+INFO:root:***** Configing Data *****
+>>> unitree_z1_stackbox: 1 data samples loaded.
+>>> unitree_z1_stackbox: data stats loaded.
+>>> unitree_z1_stackbox: normalizer initiated.
+>>> unitree_z1_dual_arm_stackbox: 1 data samples loaded.
+>>> unitree_z1_dual_arm_stackbox: data stats loaded.
+>>> unitree_z1_dual_arm_stackbox: normalizer initiated.
+>>> unitree_z1_dual_arm_stackbox_v2: 1 data samples loaded.
+>>> unitree_z1_dual_arm_stackbox_v2: data stats loaded.
+>>> unitree_z1_dual_arm_stackbox_v2: normalizer initiated.
+>>> unitree_z1_dual_arm_cleanup_pencils: 1 data samples loaded.
+>>> unitree_z1_dual_arm_cleanup_pencils: data stats loaded.
+>>> unitree_z1_dual_arm_cleanup_pencils: normalizer initiated.
+>>> unitree_g1_pack_camera: 1 data samples loaded.
+>>> unitree_g1_pack_camera: data stats loaded.
+>>> unitree_g1_pack_camera: normalizer initiated.
+>>> Dataset is successfully loaded ...
+>>> Generate 16 frames under each generation ...
+DEBUG:h5py._conv:Creating converter from 3 to 5
+DEBUG:PIL.PngImagePlugin:STREAM b'IHDR' 16 13
+DEBUG:PIL.PngImagePlugin:STREAM b'pHYs' 41 9
+DEBUG:PIL.PngImagePlugin:STREAM b'IDAT' 62 4096
+
+  0%|          | 0/8 [00:00<?, ?it/s]>>> Step 0: generating actions ...
+>>> Step 0: interacting with world model ...
+DEBUG:PIL.Image:Importing BlpImagePlugin
+DEBUG:PIL.Image:Importing BmpImagePlugin
+DEBUG:PIL.Image:Importing BufrStubImagePlugin
+DEBUG:PIL.Image:Importing CurImagePlugin
+DEBUG:PIL.Image:Importing DcxImagePlugin
+DEBUG:PIL.Image:Importing DdsImagePlugin
+DEBUG:PIL.Image:Importing EpsImagePlugin
+DEBUG:PIL.Image:Importing FitsImagePlugin
+DEBUG:PIL.Image:Importing FitsStubImagePlugin
+DEBUG:PIL.Image:Importing FliImagePlugin
+DEBUG:PIL.Image:Importing FpxImagePlugin
+DEBUG:PIL.Image:Image: failed to import FpxImagePlugin: No module named 'olefile'
+DEBUG:PIL.Image:Importing FtexImagePlugin
+DEBUG:PIL.Image:Importing GbrImagePlugin
+DEBUG:PIL.Image:Importing GifImagePlugin
+DEBUG:PIL.Image:Importing GribStubImagePlugin
+DEBUG:PIL.Image:Importing Hdf5StubImagePlugin
+DEBUG:PIL.Image:Importing IcnsImagePlugin
+DEBUG:PIL.Image:Importing IcoImagePlugin
+DEBUG:PIL.Image:Importing ImImagePlugin
+DEBUG:PIL.Image:Importing ImtImagePlugin
+DEBUG:PIL.Image:Importing IptcImagePlugin
+DEBUG:PIL.Image:Importing JpegImagePlugin
+DEBUG:PIL.Image:Importing Jpeg2KImagePlugin
+DEBUG:PIL.Image:Importing McIdasImagePlugin
+DEBUG:PIL.Image:Importing MicImagePlugin
+DEBUG:PIL.Image:Image: failed to import MicImagePlugin: No module named 'olefile'
+DEBUG:PIL.Image:Importing MpegImagePlugin
+DEBUG:PIL.Image:Importing MpoImagePlugin
+DEBUG:PIL.Image:Importing MspImagePlugin
+DEBUG:PIL.Image:Importing PalmImagePlugin
+DEBUG:PIL.Image:Importing PcdImagePlugin
+DEBUG:PIL.Image:Importing PcxImagePlugin
+DEBUG:PIL.Image:Importing PdfImagePlugin
+DEBUG:PIL.Image:Importing PixarImagePlugin
+DEBUG:PIL.Image:Importing PngImagePlugin
+DEBUG:PIL.Image:Importing PpmImagePlugin
+DEBUG:PIL.Image:Importing PsdImagePlugin
+DEBUG:PIL.Image:Importing QoiImagePlugin
+DEBUG:PIL.Image:Importing SgiImagePlugin
+DEBUG:PIL.Image:Importing SpiderImagePlugin
+DEBUG:PIL.Image:Importing SunImagePlugin
+DEBUG:PIL.Image:Importing TgaImagePlugin
+DEBUG:PIL.Image:Importing TiffImagePlugin
+DEBUG:PIL.Image:Importing WebPImagePlugin
+DEBUG:PIL.Image:Importing WmfImagePlugin
+DEBUG:PIL.Image:Importing XbmImagePlugin
+DEBUG:PIL.Image:Importing XpmImagePlugin
+DEBUG:PIL.Image:Importing XVThumbImagePlugin
+
+ 12%|█▎        | 1/8 [01:14<08:41, 74.51s/it]
+ 25%|██▌       | 2/8 [02:29<07:28, 74.79s/it]
+ 38%|███▊      | 3/8 [03:44<06:14, 74.81s/it]
+ 50%|█████     | 4/8 [04:59<04:59, 74.78s/it]
+ 62%|██████▎   | 5/8 [06:13<03:44, 74.73s/it]
+ 75%|███████▌  | 6/8 [07:28<02:29, 74.66s/it]
+ 88%|████████▊ | 7/8 [08:42<01:14, 74.56s/it]
+100%|██████████| 8/8 [09:56<00:00, 74.51s/it]
+100%|██████████| 8/8 [09:56<00:00, 74.62s/it]
+>>>>>>>>>>>>>>>>>>>>>>>>
+>>> Step 1: generating actions ...
+>>> Step 1: interacting with world model ...
+>>>>>>>>>>>>>>>>>>>>>>>>
+>>> Step 2: generating actions ...
+>>> Step 2: interacting with world model ...
+>>>>>>>>>>>>>>>>>>>>>>>>
+>>> Step 3: generating actions ...
+>>> Step 3: interacting with world model ...
+>>>>>>>>>>>>>>>>>>>>>>>>
+>>> Step 4: generating actions ...
+>>> Step 4: interacting with world model ...
+>>>>>>>>>>>>>>>>>>>>>>>>
+>>> Step 5: generating actions ...
+>>> Step 5: interacting with world model ...
+>>>>>>>>>>>>>>>>>>>>>>>>

unitree_z1_dual_arm_cleanup_pencils/case1/psnr_result.json

@@ -0,0 +1,5 @@
+{
+    "gt_video": "unitree_z1_dual_arm_cleanup_pencils/case1/unitree_z1_dual_arm_cleanup_pencils_case1.mp4",
+    "pred_video": "unitree_z1_dual_arm_cleanup_pencils/case1/output/inference/0_full_fs4.mp4",
+    "psnr": 47.911564449209735
+}

unitree_z1_dual_arm_stackbox_v2/case1/output.log

@@ -0,0 +1,120 @@
+2026-02-11 11:59:27.241485: I tensorflow/core/util/port.cc:113] oneDNN custom operations are on. You may see slightly different numerical results due to floating-point round-off errors from different computation orders. To turn them off, set the environment variable `TF_ENABLE_ONEDNN_OPTS=0`.
+2026-02-11 11:59:27.291755: E external/local_xla/xla/stream_executor/cuda/cuda_dnn.cc:9261] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered
+2026-02-11 11:59:27.291807: E external/local_xla/xla/stream_executor/cuda/cuda_fft.cc:607] Unable to register cuFFT factory: Attempting to register factory for plugin cuFFT when one has already been registered
+2026-02-11 11:59:27.293169: E external/local_xla/xla/stream_executor/cuda/cuda_blas.cc:1515] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered
+2026-02-11 11:59:27.300838: I tensorflow/core/platform/cpu_feature_guard.cc:182] This TensorFlow binary is optimized to use available CPU instructions in performance-critical operations.
+To enable the following instructions: AVX2 AVX512F AVX512_VNNI FMA, in other operations, rebuild TensorFlow with the appropriate compiler flags.
+2026-02-11 11:59:28.228009: W tensorflow/compiler/tf2tensorrt/utils/py_utils.cc:38] TF-TRT Warning: Could not find TensorRT
+Global seed set to 123
+>>> Loading prepared model from ckpts/unifolm_wma_dual.ckpt.prepared.pt ...
+>>> Prepared model loaded.
+INFO:root:***** Configing Data *****
+>>> unitree_z1_stackbox: 1 data samples loaded.
+>>> unitree_z1_stackbox: data stats loaded.
+>>> unitree_z1_stackbox: normalizer initiated.
+>>> unitree_z1_dual_arm_stackbox: 1 data samples loaded.
+>>> unitree_z1_dual_arm_stackbox: data stats loaded.
+>>> unitree_z1_dual_arm_stackbox: normalizer initiated.
+>>> unitree_z1_dual_arm_stackbox_v2: 1 data samples loaded.
+>>> unitree_z1_dual_arm_stackbox_v2: data stats loaded.
+>>> unitree_z1_dual_arm_stackbox_v2: normalizer initiated.
+>>> unitree_z1_dual_arm_cleanup_pencils: 1 data samples loaded.
+>>> unitree_z1_dual_arm_cleanup_pencils: data stats loaded.
+>>> unitree_z1_dual_arm_cleanup_pencils: normalizer initiated.
+>>> unitree_g1_pack_camera: 1 data samples loaded.
+>>> unitree_g1_pack_camera: data stats loaded.
+>>> unitree_g1_pack_camera: normalizer initiated.
+>>> Dataset is successfully loaded ...
+    ✓ KV fused: 66 attention layers
+>>> Generate 16 frames under each generation ...
+DEBUG:h5py._conv:Creating converter from 3 to 5
+DEBUG:PIL.PngImagePlugin:STREAM b'IHDR' 16 13
+DEBUG:PIL.PngImagePlugin:STREAM b'pHYs' 41 9
+DEBUG:PIL.PngImagePlugin:STREAM b'IDAT' 62 4096
+
+  0%|          | 0/11 [00:00<?, ?it/s]
+  9%|▉         | 1/11 [00:34<05:40, 34.05s/it]>>> Step 0: generating actions ...
+>>> Step 0: interacting with world model ...
+>>>>>>>>>>>>>>>>>>>>>>>>
+>>> Step 1: generating actions ...
+DEBUG:PIL.Image:Importing BlpImagePlugin
+DEBUG:PIL.Image:Importing BmpImagePlugin
+DEBUG:PIL.Image:Importing BufrStubImagePlugin
+DEBUG:PIL.Image:Importing CurImagePlugin
+DEBUG:PIL.Image:Importing DcxImagePlugin
+DEBUG:PIL.Image:Importing DdsImagePlugin
+DEBUG:PIL.Image:Importing EpsImagePlugin
+DEBUG:PIL.Image:Importing FitsImagePlugin
+DEBUG:PIL.Image:Importing FitsStubImagePlugin
+DEBUG:PIL.Image:Importing FliImagePlugin
+DEBUG:PIL.Image:Importing FpxImagePlugin
+DEBUG:PIL.Image:Image: failed to import FpxImagePlugin: No module named 'olefile'
+DEBUG:PIL.Image:Importing FtexImagePlugin
+DEBUG:PIL.Image:Importing GbrImagePlugin
+DEBUG:PIL.Image:Importing GifImagePlugin
+DEBUG:PIL.Image:Importing GribStubImagePlugin
+DEBUG:PIL.Image:Importing Hdf5StubImagePlugin
+DEBUG:PIL.Image:Importing IcnsImagePlugin
+DEBUG:PIL.Image:Importing IcoImagePlugin
+DEBUG:PIL.Image:Importing ImImagePlugin
+DEBUG:PIL.Image:Importing ImtImagePlugin
+DEBUG:PIL.Image:Importing IptcImagePlugin
+DEBUG:PIL.Image:Importing JpegImagePlugin
+DEBUG:PIL.Image:Importing Jpeg2KImagePlugin
+DEBUG:PIL.Image:Importing McIdasImagePlugin
+DEBUG:PIL.Image:Importing MicImagePlugin
+DEBUG:PIL.Image:Image: failed to import MicImagePlugin: No module named 'olefile'
+DEBUG:PIL.Image:Importing MpegImagePlugin
+DEBUG:PIL.Image:Importing MpoImagePlugin
+DEBUG:PIL.Image:Importing MspImagePlugin
+DEBUG:PIL.Image:Importing PalmImagePlugin
+DEBUG:PIL.Image:Importing PcdImagePlugin
+DEBUG:PIL.Image:Importing PcxImagePlugin
+DEBUG:PIL.Image:Importing PdfImagePlugin
+DEBUG:PIL.Image:Importing PixarImagePlugin
+DEBUG:PIL.Image:Importing PngImagePlugin
+DEBUG:PIL.Image:Importing PpmImagePlugin
+DEBUG:PIL.Image:Importing PsdImagePlugin
+DEBUG:PIL.Image:Importing QoiImagePlugin
+DEBUG:PIL.Image:Importing SgiImagePlugin
+DEBUG:PIL.Image:Importing SpiderImagePlugin
+DEBUG:PIL.Image:Importing SunImagePlugin
+DEBUG:PIL.Image:Importing TgaImagePlugin
+DEBUG:PIL.Image:Importing TiffImagePlugin
+DEBUG:PIL.Image:Importing WebPImagePlugin
+DEBUG:PIL.Image:Importing WmfImagePlugin
+DEBUG:PIL.Image:Importing XbmImagePlugin
+DEBUG:PIL.Image:Importing XpmImagePlugin
+DEBUG:PIL.Image:Importing XVThumbImagePlugin
+
+ 18%|█▊        | 2/11 [01:08<05:07, 34.17s/it]
+ 27%|██▋       | 3/11 [01:42<04:33, 34.16s/it]
+ 36%|███▋      | 4/11 [02:16<03:59, 34.18s/it]
+ 45%|████▌     | 5/11 [02:50<03:24, 34.14s/it]
+ 55%|█████▍    | 6/11 [03:24<02:50, 34.10s/it]
+ 64%|██████▎   | 7/11 [03:58<02:16, 34.07s/it]
+ 73%|███████▎  | 8/11 [04:32<01:42, 34.03s/it]
+ 82%|████████▏ | 9/11 [05:06<01:08, 34.02s/it]
+ 91%|█████████ | 10/11 [05:40<00:34, 34.04s/it]
+100%|██████████| 11/11 [06:14<00:00, 34.03s/it]
+100%|██████████| 11/11 [06:14<00:00, 34.07s/it]
+>>> Step 1: interacting with world model ...
+>>>>>>>>>>>>>>>>>>>>>>>>
+>>> Step 2: generating actions ...
+>>> Step 2: interacting with world model ...
+>>>>>>>>>>>>>>>>>>>>>>>>
+>>> Step 3: generating actions ...
+>>> Step 3: interacting with world model ...
+>>>>>>>>>>>>>>>>>>>>>>>>
+>>> Step 4: generating actions ...
+>>> Step 4: interacting with world model ...
+>>>>>>>>>>>>>>>>>>>>>>>>
+>>> Step 5: generating actions ...
+>>> Step 5: interacting with world model ...
+>>>>>>>>>>>>>>>>>>>>>>>>
+>>> Step 6: generating actions ...
+>>> Step 6: interacting with world model ...
+>>>>>>>>>>>>>>>>>>>>>>>>
+>>> Step 7: generating actions ...
+>>> Step 7: interacting with world model ...
+>>>>>>>>>>>>>>>>>>>>>>>>

unitree_z1_dual_arm_stackbox_v2/case1/psnr_result.json

@@ -0,0 +1,5 @@
+{
+    "gt_video": "/home/qhy/unifolm-world-model-action/unitree_z1_dual_arm_stackbox_v2/case1/unitree_z1_dual_arm_stackbox_v2_case1.mp4",
+    "pred_video": "/home/qhy/unifolm-world-model-action/unitree_z1_dual_arm_stackbox_v2/case1/output/inference/5_full_fs4.mp4",
+    "psnr": 28.167025381705358
+}

unitree_z1_dual_arm_stackbox_v2/case1/run_world_model_interaction.sh

@@ -20,5 +20,6 @@ dataset="unitree_z1_dual_arm_stackbox_v2"
         --n_iter 11 \
         --timestep_spacing 'uniform_trailing' \
         --guidance_rescale 0.7 \
-        --perframe_ae
+        --perframe_ae \
+        --fast_policy_no_decode
 } 2>&1 | tee "${res_dir}/output.log"