Add Aloha env and ACT policy

WIP Aloha env tests pass

Rendering works (fps look fast tho? TODO action bounding is too wide [-1,1])

Update README

Copy past from act repo

Remove download.py add a WIP for Simxarm

Remove download.py add a WIP for Simxarm

Add act yaml (TODO: try train.py)

Training can runs (TODO: eval)

Add tasks without end_effector that are compatible with dataset, Eval can run (TODO: training and pretrained model)

Add AbstractEnv, Refactor AlohaEnv, Add rendering_hook in env, Minor modifications, (TODO: Refactor Pusht and Simxarm)

poetry lock

fix bug in compute_stats for action normalization

fix more bugs in normalization

fix training

fix import

PushtEnv inheriates AbstractEnv, Improve factory Normalization

Add _make_env to EnvAbstract

Add call_rendering_hooks to pusht env

SimxarmEnv inherites from AbstractEnv (NOT TESTED)

Add aloha tests artifacts + update pusht stats

fix image normalization: before env was in [0,1] but dataset in [0,255], and now both in [0,255]

Small fix on simxarm

Add next to obs

Add top camera to Aloha env (TODO: make it compatible with set of cameras)

Add top camera to Aloha env (TODO: make it compatible with set of cameras)

lerobot/common/policies/act/position_encoding.py

@@ -0,0 +1,101 @@
+"""
+Various positional encodings for the transformer.
+"""
+import math
+
+import torch
+from torch import nn
+
+from .utils import NestedTensor
+
+
+class PositionEmbeddingSine(nn.Module):
+    """
+    This is a more standard version of the position embedding, very similar to the one
+    used by the Attention is all you need paper, generalized to work on images.
+    """
+
+    def __init__(self, num_pos_feats=64, temperature=10000, normalize=False, scale=None):
+        super().__init__()
+        self.num_pos_feats = num_pos_feats
+        self.temperature = temperature
+        self.normalize = normalize
+        if scale is not None and normalize is False:
+            raise ValueError("normalize should be True if scale is passed")
+        if scale is None:
+            scale = 2 * math.pi
+        self.scale = scale
+
+    def forward(self, tensor):
+        x = tensor
+        # mask = tensor_list.mask
+        # assert mask is not None
+        # not_mask = ~mask
+
+        not_mask = torch.ones_like(x[0, [0]])
+        y_embed = not_mask.cumsum(1, dtype=torch.float32)
+        x_embed = not_mask.cumsum(2, dtype=torch.float32)
+        if self.normalize:
+            eps = 1e-6
+            y_embed = y_embed / (y_embed[:, -1:, :] + eps) * self.scale
+            x_embed = x_embed / (x_embed[:, :, -1:] + eps) * self.scale
+
+        dim_t = torch.arange(self.num_pos_feats, dtype=torch.float32, device=x.device)
+        dim_t = self.temperature ** (2 * (dim_t // 2) / self.num_pos_feats)
+
+        pos_x = x_embed[:, :, :, None] / dim_t
+        pos_y = y_embed[:, :, :, None] / dim_t
+        pos_x = torch.stack((pos_x[:, :, :, 0::2].sin(), pos_x[:, :, :, 1::2].cos()), dim=4).flatten(3)
+        pos_y = torch.stack((pos_y[:, :, :, 0::2].sin(), pos_y[:, :, :, 1::2].cos()), dim=4).flatten(3)
+        pos = torch.cat((pos_y, pos_x), dim=3).permute(0, 3, 1, 2)
+        return pos
+
+
+class PositionEmbeddingLearned(nn.Module):
+    """
+    Absolute pos embedding, learned.
+    """
+
+    def __init__(self, num_pos_feats=256):
+        super().__init__()
+        self.row_embed = nn.Embedding(50, num_pos_feats)
+        self.col_embed = nn.Embedding(50, num_pos_feats)
+        self.reset_parameters()
+
+    def reset_parameters(self):
+        nn.init.uniform_(self.row_embed.weight)
+        nn.init.uniform_(self.col_embed.weight)
+
+    def forward(self, tensor_list: NestedTensor):
+        x = tensor_list.tensors
+        h, w = x.shape[-2:]
+        i = torch.arange(w, device=x.device)
+        j = torch.arange(h, device=x.device)
+        x_emb = self.col_embed(i)
+        y_emb = self.row_embed(j)
+        pos = (
+            torch.cat(
+                [
+                    x_emb.unsqueeze(0).repeat(h, 1, 1),
+                    y_emb.unsqueeze(1).repeat(1, w, 1),
+                ],
+                dim=-1,
+            )
+            .permute(2, 0, 1)
+            .unsqueeze(0)
+            .repeat(x.shape[0], 1, 1, 1)
+        )
+        return pos
+
+
+def build_position_encoding(args):
+    n_steps = args.hidden_dim // 2
+    if args.position_embedding in ("v2", "sine"):
+        # TODO find a better way of exposing other arguments
+        position_embedding = PositionEmbeddingSine(n_steps, normalize=True)
+    elif args.position_embedding in ("v3", "learned"):
+        position_embedding = PositionEmbeddingLearned(n_steps)
+    else:
+        raise ValueError(f"not supported {args.position_embedding}")
+
+    return position_embedding