From 42fbcc89c5d7c7d54686c17498e0167562d89c39 Mon Sep 17 00:00:00 2001 From: Pepijn Date: Sun, 31 Aug 2025 02:10:52 +0200 Subject: [PATCH] ddebugging --- notebooks/rlearn_evaluation.ipynb | 47 ++++----- .../policies/rlearn/modeling_rlearn.py | 99 ++++++++++++++----- 2 files changed, 97 insertions(+), 49 deletions(-) diff --git a/notebooks/rlearn_evaluation.ipynb b/notebooks/rlearn_evaluation.ipynb index 630726db8..0618245cf 100644 --- a/notebooks/rlearn_evaluation.ipynb +++ b/notebooks/rlearn_evaluation.ipynb @@ -234,28 +234,21 @@ }, { "cell_type": "code", - "execution_count": 6, + "execution_count": 7, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ - "Visualizing first 4 episodes...\n" - ] - }, - { - "name": "stderr", - "output_type": "stream", - "text": [ - "objc[8231]: Class AVFFrameReceiver is implemented in both /Users/pepijnkooijmans/miniconda3/envs/lerobot/lib/python3.10/site-packages/av/.dylibs/libavdevice.61.3.100.dylib (0x1696703a8) and /Users/pepijnkooijmans/miniconda3/envs/lerobot/lib/libavdevice.61.3.100.dylib (0x30a620848). This may cause spurious casting failures and mysterious crashes. One of the duplicates must be removed or renamed.\n", - "objc[8231]: Class AVFAudioReceiver is implemented in both /Users/pepijnkooijmans/miniconda3/envs/lerobot/lib/python3.10/site-packages/av/.dylibs/libavdevice.61.3.100.dylib (0x1696703f8) and /Users/pepijnkooijmans/miniconda3/envs/lerobot/lib/libavdevice.61.3.100.dylib (0x30a620898). This may cause spurious casting failures and mysterious crashes. One of the duplicates must be removed or renamed.\n", - "W0831 01:34:23.580000 8231 site-packages/torch/_dynamo/variables/builtin.py:861] [0/0] incorrect arg count too many positional arguments and no constant handler\n" + "Visualizing first 4 episodes...\n", + "Error processing episode 0: The size of tensor a (128) must match the size of tensor b (16) at non-singleton dimension 1\n", + "Error processing episode 1: The size of tensor a (128) must match the size of tensor b (16) at non-singleton dimension 1\n" ] }, { "data": { - "image/png": "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", + "image/png": "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", "text/plain": [ "
" ] @@ -270,20 +263,10 @@ "\n", "Detailed Episode Statistics:\n", "================================================================================\n", - "Episode 0:\n", - " Language: Pickup the blue block\n", - " Frames: 128\n", - " VOC-S (Spearman ρ): 0.1152 (p=0.1955)\n", - " Reward range: [0.533, 0.545]\n", - " Trend: ↗ Increasing\n", - "--------------------------------------------------------------------------------\n", - "Episode 1:\n", - " Language: Pickup the blue block\n", - " Frames: 128\n", - " VOC-S (Spearman ρ): 0.0775 (p=0.3846)\n", - " Reward range: [0.535, 0.545]\n", - " Trend: → Flat\n", - "--------------------------------------------------------------------------------\n" + "Error processing episode 0: The size of tensor a (128) must match the size of tensor b (16) at non-singleton dimension 1\n", + "Episode 0: No data available\n", + "Error processing episode 1: The size of tensor a (128) must match the size of tensor b (16) at non-singleton dimension 1\n", + "Episode 1: No data available\n" ] } ], @@ -353,7 +336,17 @@ " frames_tensor = torch.stack(frames)\n", "\n", " # Predict rewards\n", - " rewards = evaluator.predict_episode_rewards(frames_tensor, language)\n", + " # CORRECTED: Use current model architecture\n", + " # Create batch dict in the format expected by predict_rewards\n", + " batch = {\n", + " \"observation.images\": frames_tensor.unsqueeze(0), # (1, T, C, H, W)\n", + " \"observation.language\": [language] # List of strings\n", + " }\n", + "\n", + " # Use the model's predict_rewards method directly\n", + " with torch.no_grad():\n", + " rewards_tensor = model.predict_rewards(batch) # (1, T)\n", + " rewards = rewards_tensor.squeeze(0).cpu().numpy() # (T,)\n", "\n", " return frames_tensor, language, rewards, episode_length\n", "\n", diff --git a/src/lerobot/policies/rlearn/modeling_rlearn.py b/src/lerobot/policies/rlearn/modeling_rlearn.py index ea40aad1a..ad1c448ee 100644 --- a/src/lerobot/policies/rlearn/modeling_rlearn.py +++ b/src/lerobot/policies/rlearn/modeling_rlearn.py @@ -507,17 +507,39 @@ class RLearNPolicy(PreTrainedPolicy): # Calculate progress for each frame in the temporal window all_progress = [] - # DEBUG: Log first sample's target calculation - debug_first_sample = True - if debug_first_sample and torch.rand(1).item() < 0.05: # 5% chance - ep_idx_debug = episode_indices[0].item() - frame_idx_debug = frame_indices[0].item() - ep_start_debug = self.episode_data_index["from"][ep_idx_debug].item() - ep_end_debug = self.episode_data_index["to"][ep_idx_debug].item() - ep_length_debug = ep_end_debug - ep_start_debug - print(f"\n=== TARGET DEBUG ===") - print(f"Episode {ep_idx_debug}: length={ep_length_debug}, current_frame={frame_idx_debug}") + # DEBUG: Log indexing details for first sample occasionally + debug_indexing = torch.rand(1).item() < 0.05 # 5% chance + if debug_indexing: + print(f"\n=== INDEXING DEBUG ===") print(f"Delta indices: {delta_indices}") + print(f"Batch size: {B}") + + # Check if batch samples have diverse frame indices (red flag if all identical) + unique_frames = torch.unique(frame_indices).tolist() + unique_episodes = torch.unique(episode_indices).tolist() + print(f"Unique frame indices in batch: {unique_frames[:10]}{'...' if len(unique_frames) > 10 else ''}") + print(f"Unique episode indices in batch: {unique_episodes[:10]}{'...' if len(unique_episodes) > 10 else ''}") + + if len(unique_frames) == 1: + print("🚨 RED FLAG: All samples have IDENTICAL frame index! This causes identical targets.") + + # First sample details + ep_idx_0 = episode_indices[0].item() + frame_idx_0 = frame_indices[0].item() + ep_start_0 = self.episode_data_index["from"][ep_idx_0].item() + ep_end_0 = self.episode_data_index["to"][ep_idx_0].item() + ep_length_0 = ep_end_0 - ep_start_0 + print(f"First sample - Episode: {ep_idx_0}, Frame: {frame_idx_0}/{ep_length_0}, Episode length: {ep_length_0}") + + # Check boundary proximity + frames_from_start = frame_idx_0 + frames_from_end = ep_length_0 - frame_idx_0 - 1 + print(f"First sample proximity - Start: {frames_from_start}, End: {frames_from_end}") + + if frames_from_start < 15: + print(f"⚠️ Close to episode START: many deltas will go negative") + if frames_from_end < 15: + print(f"⚠️ Close to episode END: many deltas will exceed episode") for i, delta in enumerate(delta_indices): # For each sample, calculate the progress of the frame at delta offset @@ -534,23 +556,32 @@ class RLearNPolicy(PreTrainedPolicy): ep_end = self.episode_data_index["to"][ep_idx].item() ep_length = ep_end - ep_start - # Clamp to episode boundaries (frame_index is relative to episode) - target_frame_idx_clamped = max(0, min(ep_length - 1, target_frame_idx)) - - # Calculate progress for this frame - prog = target_frame_idx_clamped / max(1, ep_length - 1) + # Calculate progress with proper boundary handling + if target_frame_idx < 0: + # Before episode start: extrapolate negative progress + prog = target_frame_idx / max(1, ep_length - 1) + elif target_frame_idx >= ep_length: + # After episode end: extrapolate progress beyond 1.0 + prog = target_frame_idx / max(1, ep_length - 1) + else: + # Within episode: normal progress calculation + prog = target_frame_idx / max(1, ep_length - 1) + + # Clip to reasonable bounds to prevent extreme values + prog = max(-1.0, min(2.0, prog)) # Allow some extrapolation frame_progress.append(prog) - # DEBUG: Log first sample calculation - if debug_first_sample and b_idx == 0 and torch.rand(1).item() < 0.05: - print(f"Frame {i:2d} (delta={delta:3d}): target_idx={target_frame_idx:3d} → clamped={target_frame_idx_clamped:3d} → progress={prog:.6f}") + # DEBUG: Log first sample's calculation + if debug_indexing and b_idx == 0: + boundary_status = "BEFORE" if target_frame_idx < 0 else "AFTER" if target_frame_idx >= ep_length else "WITHIN" + print(f" Frame {i:2d} (δ={delta:3d}): target_idx={target_frame_idx:3d} [{boundary_status}] → progress={prog:.6f}") all_progress.append( torch.tensor(frame_progress, device=video_frame_embeds.device, dtype=video_frame_embeds.dtype) ) - - if debug_first_sample and torch.rand(1).item() < 0.05: - print("=" * 20) + + if debug_indexing: + print("=" * 22) # Stack to get (B, T) tensor where T is the temporal sequence length target = torch.stack(all_progress, dim=1) # (B, max_seq_len) @@ -650,6 +681,23 @@ class RLearNPolicy(PreTrainedPolicy): # Show randomly sampled sequence for comparison print(f"Sample {sample_idx} targets (all 16):", target[sample_idx].cpu().numpy()) print(f"Sample {sample_idx} preds (all 16): ", preds[sample_idx].cpu().numpy()) + + # TARGET FIX VERIFICATION: Check if we still have flat/stuck patterns + sample_targets = target[sample_idx].cpu().numpy() + # Count consecutive identical values (should be minimal after fix) + consecutive_same = 0 + max_consecutive = 0 + for i in range(1, len(sample_targets)): + if abs(sample_targets[i] - sample_targets[i-1]) < 1e-6: + consecutive_same += 1 + max_consecutive = max(max_consecutive, consecutive_same + 1) + else: + consecutive_same = 0 + + if max_consecutive >= 3: + print(f"⚠️ STILL STUCK: {max_consecutive} consecutive identical targets!") + else: + print(f"✅ TARGET FIXED: Max consecutive identical = {max_consecutive}") print("="*25) total_forward_time = time.perf_counter() - forward_start @@ -912,7 +960,7 @@ def extract_visual_sequence(batch: dict[str, Tensor], target_seq_len: int = None f"All keys: {available_keys}" ) - # Pad sequence if needed + # Adjust sequence length if needed if target_seq_len is not None: B, T, C, H, W = frames.shape if T < target_seq_len: @@ -925,6 +973,13 @@ def extract_visual_sequence(batch: dict[str, Tensor], target_seq_len: int = None import logging logging.debug(f"Padded sequence from {T} to {target_seq_len} frames by repeating first frame") + elif T > target_seq_len: + # Truncate to target length, keeping the most recent frames + frames = frames[:, -target_seq_len:] + + import logging + + logging.debug(f"Truncated sequence from {T} to {target_seq_len} frames by keeping most recent frames") return frames