perf(pi0fast): stop FAST decoding at the end-of-action marker

The decode loop always runs all max_decoding_steps (256) tokens, but
detokenize_actions() cuts the output at the first "|", so most of them get
generated and then thrown away. On LIBERO the marker lands around token 25-35.

Stopping there gives the same actions ~4x faster end to end. Checked on
libero_10: 50/50 episodes came out identical to the old path, same success
and same state trajectories.

src/lerobot/policies/pi0_fast/modeling_pi0_fast.py

@@ -612,6 +612,11 @@ class PI0FastPytorch(nn.Module):  # see openpi `PI0Pytorch`
         device = tokens.device
         lm_head = self.paligemma_with_expert.paligemma.lm_head
 
+        # detokenize_actions() cuts at the first "|", so anything generated past it is
+        # discarded anyway.
+        pipe_id = self._paligemma_tokenizer.convert_tokens_to_ids("|")
+        finished = torch.zeros(bsize, dtype=torch.bool, device=device)
+
         # --- 1. PREFILL PHASE ---
         # Process Images + Text Prompt + BOS token once to populate the KV cache.
 
@@ -663,6 +668,7 @@ class PI0FastPytorch(nn.Module):  # see openpi `PI0Pytorch`
         # Initialize storage for generated tokens
         generated_action_tokens = torch.zeros((bsize, max_decoding_steps), dtype=torch.long, device=device)
         generated_action_tokens[:, 0] = next_token.squeeze(-1)
+        finished |= next_token.squeeze(-1) == pipe_id
 
         # Track valid tokens mask (0 for pad, 1 for valid)
         # We need this to tell the new token what it can attend to (images + text + past actions)
@@ -713,6 +719,10 @@ class PI0FastPytorch(nn.Module):  # see openpi `PI0Pytorch`
 
             generated_action_tokens[:, t] = next_token.squeeze(-1)
 
+            finished |= next_token.squeeze(-1) == pipe_id
+            if bool(finished.all()):
+                break
+
         return generated_action_tokens