Upload folder using huggingface_hub
79f4ad1 verified about 2 months ago
31.5 kB
[INFO] Running in WANDB offline mode
[INFO] Running in WANDB offline mode
[INFO] Running in WANDB offline mode
[INFO] Running in WANDB offline mode
Trainer._get_train_sampler replaced with custom implementation.
warmup_ratio is deprecated and will be removed in v5.2. Use `warmup_steps` instead.
Trainer._get_train_sampler replaced with custom implementation.
Trainer._get_train_sampler replaced with custom implementation.
warmup_ratio is deprecated and will be removed in v5.2. Use `warmup_steps` instead.
[2026-09-18 05:18:27,633] [INFO] [real_accelerator.py:222:get_accelerator] Setting ds_accelerator to cuda (auto detect)
warmup_ratio is deprecated and will be removed in v5.2. Use `warmup_steps` instead.
[2026-09-18 05:18:27,660] [INFO] [real_accelerator.py:222:get_accelerator] Setting ds_accelerator to cuda (auto detect)
[2026-09-18 05:18:27,668] [INFO] [real_accelerator.py:222:get_accelerator] Setting ds_accelerator to cuda (auto detect)
Trainer._get_train_sampler replaced with custom implementation.
warmup_ratio is deprecated and will be removed in v5.2. Use `warmup_steps` instead.
[2026-09-18 05:18:27,847] [INFO] [real_accelerator.py:222:get_accelerator] Setting ds_accelerator to cuda (auto detect)
[2026-09-18 05:18:27,972] [INFO] [comm.py:658:init_distributed] cdb=None
[2026-09-18 05:18:27,972] [INFO] [comm.py:658:init_distributed] cdb=None
[2026-09-18 05:18:27,995] [INFO] [comm.py:658:init_distributed] cdb=None
[2026-09-18 05:18:28,162] [INFO] [comm.py:658:init_distributed] cdb=None
[2026-09-18 05:18:28,162] [INFO] [comm.py:689:init_distributed] Initializing TorchBackend in DeepSpeed with backend nccl
[TrainingFix] Qwen3.5 FlashAttention MRoPE packing guard enabled
[TrainingFix] Qwen3.5 FlashAttention MRoPE packing guard enabled
[TrainingFix] Qwen3.5 FlashAttention MRoPE packing guard enabled
[TrainingFix] Qwen3.5 FlashAttention MRoPE packing guard enabled

Loading weights:   0%|          | 0/723 [00:00<?, ?it/s]
Loading weights:   0%|          | 0/723 [00:00<?, ?it/s]
Loading weights: 100%|██████████| 723/723 [00:00<00:00, 12793.99it/s]

Loading weights:   0%|          | 0/723 [00:00<?, ?it/s]
Loading weights: 100%|██████████| 723/723 [00:00<00:00, 11519.09it/s]

Loading weights:   0%|          | 0/723 [00:00<?, ?it/s]
Loading weights: 100%|██████████| 723/723 [00:00<00:00, 11001.00it/s]

Loading weights: 100%|██████████| 723/723 [00:00<00:00, 11823.92it/s]
[VGGTEncoder] eval KV-cache window: start=8 recent=48 (total=56 frames)
[FeatureFusion] geometry fusion delta scaled by 0.5
[FeatureFusion] importance gate ENABLED on geometry delta
[FeatureFusion] learnable geometry scale ENABLED (init 0.5)
[VGGTEncoder] eval KV-cache window: start=8 recent=48 (total=56 frames)
[FeatureFusion] geometry fusion delta scaled by 0.5
[FeatureFusion] importance gate ENABLED on geometry delta
[FeatureFusion] learnable geometry scale ENABLED (init 0.5)
[VGGTEncoder] eval KV-cache window: start=8 recent=48 (total=56 frames)
[FeatureFusion] geometry fusion delta scaled by 0.5
[FeatureFusion] importance gate ENABLED on geometry delta
[FeatureFusion] learnable geometry scale ENABLED (init 0.5)
[VGGTEncoder] eval KV-cache window: start=8 recent=48 (total=56 frames)
[FeatureFusion] geometry fusion delta scaled by 0.5
[FeatureFusion] importance gate ENABLED on geometry delta
[FeatureFusion] learnable geometry scale ENABLED (init 0.5)
Loading weights from local directory
Loading weights from local directory
Loading weights from local directory
Loading weights from local directory
>>>>> STOP loss up-weight: 3.0 on token ids [48964]
Vision Module - Attention Blocks:
Trainable Block Indices: None
Non-Trainable Block Indices: [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23]
Merger Module Trainable: True
LLM Module - Embed Tokens Trainable: True
LLM Module - Trainable Layer Indices: [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31]
LLM Module - Non-Trainable Layer Indices: None
Qwen3_5Config {
  "architectures": [
    "Qwen3_5ForConditionalGeneration"
  ],
  "dtype": "bfloat16",
  "feature_fusion_method": "deepstack_language_add",
  "fusion_num_layers": 1,
  "geometry_encoder_layers": [
    11,
    17,
    23
  ],
  "geometry_encoder_path": "/mnt/data/vmo-ai-task/dungpq6/model-checkpoint/VGGT-1B",
  "geometry_encoder_streaming": true,
  "geometry_encoder_type": "vggt",
  "geometry_frame_strict": true,
  "geometry_fusion_layers": [
    0,
    1,
    2
  ],
  "geometry_fusion_scale": 0.5,
  "geometry_importance_gate": true,
  "geometry_learnable_scale": true,
  "geometry_merger_type": "mlp",
  "geometry_spatial_bias": false,
  "image_token_id": 248056,
  "include_camera_token": false,
  "model_type": "qwen3_5",
  "pos_encoding_type": "none",
  "reference_frame": "first",
  "supervised_token_loss": true,
  "text_config": {
    "attention_bias": false,
    "attention_dropout": 0.0,
    "attn_output_gate": true,
    "bos_token_id": null,
    "dtype": "bfloat16",
    "eos_token_id": 248044,
    "full_attention_interval": 4,
    "head_dim": 256,
    "hidden_act": "silu",
    "hidden_size": 2560,
    "initializer_range": 0.02,
    "intermediate_size": 9216,
    "layer_types": [
      "linear_attention",
      "linear_attention",
      "linear_attention",
      "full_attention",
      "linear_attention",
      "linear_attention",
      "linear_attention",
      "full_attention",
      "linear_attention",
      "linear_attention",
      "linear_attention",
      "full_attention",
      "linear_attention",
      "linear_attention",
      "linear_attention",
      "full_attention",
      "linear_attention",
      "linear_attention",
      "linear_attention",
      "full_attention",
      "linear_attention",
      "linear_attention",
      "linear_attention",
      "full_attention",
      "linear_attention",
      "linear_attention",
      "linear_attention",
      "full_attention",
      "linear_attention",
      "linear_attention",
      "linear_attention",
      "full_attention"
    ],
    "linear_conv_kernel_dim": 4,
    "linear_key_head_dim": 128,
    "linear_num_key_heads": 16,
    "linear_num_value_heads": 32,
    "linear_value_head_dim": 128,
    "mamba_ssm_dtype": "float32",
    "max_position_embeddings": 262144,
    "mlp_only_layers": [],
    "model_type": "qwen3_5_text",
    "mtp_num_hidden_layers": 1,
    "mtp_use_dedicated_embeddings": false,
    "num_attention_heads": 16,
    "num_hidden_layers": 32,
    "num_key_value_heads": 4,
    "pad_token_id": null,
    "partial_rotary_factor": 0.25,
    "rms_norm_eps": 1e-06,
    "rope_parameters": {
      "mrope_interleaved": true,
      "mrope_section": [
        11,
        11,
        10
      ],
      "partial_rotary_factor": 0.25,
      "rope_theta": 10000000,
      "rope_type": "default"
    },
    "tie_word_embeddings": true,
    "use_cache": true,
    "vocab_size": 248320
  },
  "tie_word_embeddings": true,
  "transformers_version": "5.3.0",
  "use_cache": false,
  "use_geometry_encoder": true,
  "video_token_id": 248057,
  "vision_config": {
    "deepstack_visual_indexes": [],
    "depth": 24,
    "dtype": "bfloat16",
    "hidden_act": "gelu_pytorch_tanh",
    "hidden_size": 1024,
    "in_channels": 3,
    "initializer_range": 0.02,
    "intermediate_size": 4096,
    "model_type": "qwen3_5",
    "num_heads": 16,
    "num_position_embeddings": 2304,
    "out_hidden_size": 2560,
    "patch_size": 16,
    "spatial_merge_size": 2,
    "temporal_patch_size": 2
  },
  "vision_end_token_id": 248054,
  "vision_language_fusion_layers": null,
  "vision_start_token_id": 248053
}

Loading datasets: [{'annotation_path': '/mnt/data/vmo-ai-task/anhdh35/JanusVLN/train_r2r_rxr.json', 'data_path': '/mnt/data/vmo-ai-task/anhdh35/JanusVLN', 'tag': '3d', 'sampling_rate': 1.0, 'dataset_name': 'train_r2r_rxr'}]
>>>>> STOP loss up-weight: 3.0 on token ids [48964]
Qwen3_5Config {
  "architectures": [
    "Qwen3_5ForConditionalGeneration"
  ],
  "dtype": "bfloat16",
  "feature_fusion_method": "deepstack_language_add",
  "fusion_num_layers": 1,
  "geometry_encoder_layers": [
    11,
    17,
    23
  ],
  "geometry_encoder_path": "/mnt/data/vmo-ai-task/dungpq6/model-checkpoint/VGGT-1B",
  "geometry_encoder_streaming": true,
  "geometry_encoder_type": "vggt",
  "geometry_frame_strict": true,
  "geometry_fusion_layers": [
    0,
    1,
    2
  ],
  "geometry_fusion_scale": 0.5,
  "geometry_importance_gate": true,
  "geometry_learnable_scale": true,
  "geometry_merger_type": "mlp",
  "geometry_spatial_bias": false,
  "image_token_id": 248056,
  "include_camera_token": false,
  "model_type": "qwen3_5",
  "pos_encoding_type": "none",
  "reference_frame": "first",
  "supervised_token_loss": true,
  "text_config": {
    "attention_bias": false,
    "attention_dropout": 0.0,
    "attn_output_gate": true,
    "bos_token_id": null,
    "dtype": "bfloat16",
    "eos_token_id": 248044,
    "full_attention_interval": 4,
    "head_dim": 256,
    "hidden_act": "silu",
    "hidden_size": 2560,
    "initializer_range": 0.02,
    "intermediate_size": 9216,
    "layer_types": [
      "linear_attention",
      "linear_attention",
      "linear_attention",
      "full_attention",
      "linear_attention",
      "linear_attention",
      "linear_attention",
      "full_attention",
      "linear_attention",
      "linear_attention",
      "linear_attention",
      "full_attention",
      "linear_attention",
      "linear_attention",
      "linear_attention",
      "full_attention",
      "linear_attention",
      "linear_attention",
      "linear_attention",
      "full_attention",
      "linear_attention",
      "linear_attention",
      "linear_attention",
      "full_attention",
      "linear_attention",
      "linear_attention",
      "linear_attention",
      "full_attention",
      "linear_attention",
      "linear_attention",
      "linear_attention",
      "full_attention"
    ],
    "linear_conv_kernel_dim": 4,
    "linear_key_head_dim": 128,
    "linear_num_key_heads": 16,
    "linear_num_value_heads": 32,
    "linear_value_head_dim": 128,
    "mamba_ssm_dtype": "float32",
    "max_position_embeddings": 262144,
    "mlp_only_layers": [],
    "model_type": "qwen3_5_text",
    "mtp_num_hidden_layers": 1,
    "mtp_use_dedicated_embeddings": false,
    "num_attention_heads": 16,
    "num_hidden_layers": 32,
    "num_key_value_heads": 4,
    "pad_token_id": null,
    "partial_rotary_factor": 0.25,
    "rms_norm_eps": 1e-06,
    "rope_parameters": {
      "mrope_interleaved": true,
      "mrope_section": [
        11,
        11,
        10
      ],
      "partial_rotary_factor": 0.25,
      "rope_theta": 10000000,
      "rope_type": "default"
    },
    "tie_word_embeddings": true,
    "use_cache": true,
    "vocab_size": 248320
  },
  "tie_word_embeddings": true,
  "transformers_version": "5.3.0",
  "use_cache": false,
  "use_geometry_encoder": true,
  "video_token_id": 248057,
  "vision_config": {
    "deepstack_visual_indexes": [],
    "depth": 24,
    "dtype": "bfloat16",
    "hidden_act": "gelu_pytorch_tanh",
    "hidden_size": 1024,
    "in_channels": 3,
    "initializer_range": 0.02,
    "intermediate_size": 4096,
    "model_type": "qwen3_5",
    "num_heads": 16,
    "num_position_embeddings": 2304,
    "out_hidden_size": 2560,
    "patch_size": 16,
    "spatial_merge_size": 2,
    "temporal_patch_size": 2
  },
  "vision_end_token_id": 248054,
  "vision_language_fusion_layers": null,
  "vision_start_token_id": 248053
}

Loading datasets: [{'annotation_path': '/mnt/data/vmo-ai-task/anhdh35/JanusVLN/train_r2r_rxr.json', 'data_path': '/mnt/data/vmo-ai-task/anhdh35/JanusVLN', 'tag': '3d', 'sampling_rate': 1.0, 'dataset_name': 'train_r2r_rxr'}]
>>>>> STOP loss up-weight: 3.0 on token ids [48964]
Qwen3_5Config {
  "architectures": [
    "Qwen3_5ForConditionalGeneration"
  ],
  "dtype": "bfloat16",
  "feature_fusion_method": "deepstack_language_add",
  "fusion_num_layers": 1,
  "geometry_encoder_layers": [
    11,
    17,
    23
  ],
  "geometry_encoder_path": "/mnt/data/vmo-ai-task/dungpq6/model-checkpoint/VGGT-1B",
  "geometry_encoder_streaming": true,
  "geometry_encoder_type": "vggt",
  "geometry_frame_strict": true,
  "geometry_fusion_layers": [
    0,
    1,
    2
  ],
  "geometry_fusion_scale": 0.5,
  "geometry_importance_gate": true,
  "geometry_learnable_scale": true,
  "geometry_merger_type": "mlp",
  "geometry_spatial_bias": false,
  "image_token_id": 248056,
  "include_camera_token": false,
  "model_type": "qwen3_5",
  "pos_encoding_type": "none",
  "reference_frame": "first",
  "supervised_token_loss": true,
  "text_config": {
    "attention_bias": false,
    "attention_dropout": 0.0,
    "attn_output_gate": true,
    "bos_token_id": null,
    "dtype": "bfloat16",
    "eos_token_id": 248044,
    "full_attention_interval": 4,
    "head_dim": 256,
    "hidden_act": "silu",
    "hidden_size": 2560,
    "initializer_range": 0.02,
    "intermediate_size": 9216,
    "layer_types": [
      "linear_attention",
      "linear_attention",
      "linear_attention",
      "full_attention",
      "linear_attention",
      "linear_attention",
      "linear_attention",
      "full_attention",
      "linear_attention",
      "linear_attention",
      "linear_attention",
      "full_attention",
      "linear_attention",
      "linear_attention",
      "linear_attention",
      "full_attention",
      "linear_attention",
      "linear_attention",
      "linear_attention",
      "full_attention",
      "linear_attention",
      "linear_attention",
      "linear_attention",
      "full_attention",
      "linear_attention",
      "linear_attention",
      "linear_attention",
      "full_attention",
      "linear_attention",
      "linear_attention",
      "linear_attention",
      "full_attention"
    ],
    "linear_conv_kernel_dim": 4,
    "linear_key_head_dim": 128,
    "linear_num_key_heads": 16,
    "linear_num_value_heads": 32,
    "linear_value_head_dim": 128,
    "mamba_ssm_dtype": "float32",
    "max_position_embeddings": 262144,
    "mlp_only_layers": [],
    "model_type": "qwen3_5_text",
    "mtp_num_hidden_layers": 1,
    "mtp_use_dedicated_embeddings": false,
    "num_attention_heads": 16,
    "num_hidden_layers": 32,
    "num_key_value_heads": 4,
    "pad_token_id": null,
    "partial_rotary_factor": 0.25,
    "rms_norm_eps": 1e-06,
    "rope_parameters": {
      "mrope_interleaved": true,
      "mrope_section": [
        11,
        11,
        10
      ],
      "partial_rotary_factor": 0.25,
      "rope_theta": 10000000,
      "rope_type": "default"
    },
    "tie_word_embeddings": true,
    "use_cache": true,
    "vocab_size": 248320
  },
  "tie_word_embeddings": true,
  "transformers_version": "5.3.0",
  "use_cache": false,
  "use_geometry_encoder": true,
  "video_token_id": 248057,
  "vision_config": {
    "deepstack_visual_indexes": [],
    "depth": 24,
    "dtype": "bfloat16",
    "hidden_act": "gelu_pytorch_tanh",
    "hidden_size": 1024,
    "in_channels": 3,
    "initializer_range": 0.02,
    "intermediate_size": 4096,
    "model_type": "qwen3_5",
    "num_heads": 16,
    "num_position_embeddings": 2304,
    "out_hidden_size": 2560,
    "patch_size": 16,
    "spatial_merge_size": 2,
    "temporal_patch_size": 2
  },
  "vision_end_token_id": 248054,
  "vision_language_fusion_layers": null,
  "vision_start_token_id": 248053
}

Loading datasets: [{'annotation_path': '/mnt/data/vmo-ai-task/anhdh35/JanusVLN/train_r2r_rxr.json', 'data_path': '/mnt/data/vmo-ai-task/anhdh35/JanusVLN', 'tag': '3d', 'sampling_rate': 1.0, 'dataset_name': 'train_r2r_rxr'}]
>>>>> STOP loss up-weight: 3.0 on token ids [48964]
Qwen3_5Config {
  "architectures": [
    "Qwen3_5ForConditionalGeneration"
  ],
  "dtype": "bfloat16",
  "feature_fusion_method": "deepstack_language_add",
  "fusion_num_layers": 1,
  "geometry_encoder_layers": [
    11,
    17,
    23
  ],
  "geometry_encoder_path": "/mnt/data/vmo-ai-task/dungpq6/model-checkpoint/VGGT-1B",
  "geometry_encoder_streaming": true,
  "geometry_encoder_type": "vggt",
  "geometry_frame_strict": true,
  "geometry_fusion_layers": [
    0,
    1,
    2
  ],
  "geometry_fusion_scale": 0.5,
  "geometry_importance_gate": true,
  "geometry_learnable_scale": true,
  "geometry_merger_type": "mlp",
  "geometry_spatial_bias": false,
  "image_token_id": 248056,
  "include_camera_token": false,
  "model_type": "qwen3_5",
  "pos_encoding_type": "none",
  "reference_frame": "first",
  "supervised_token_loss": true,
  "text_config": {
    "attention_bias": false,
    "attention_dropout": 0.0,
    "attn_output_gate": true,
    "bos_token_id": null,
    "dtype": "bfloat16",
    "eos_token_id": 248044,
    "full_attention_interval": 4,
    "head_dim": 256,
    "hidden_act": "silu",
    "hidden_size": 2560,
    "initializer_range": 0.02,
    "intermediate_size": 9216,
    "layer_types": [
      "linear_attention",
      "linear_attention",
      "linear_attention",
      "full_attention",
      "linear_attention",
      "linear_attention",
      "linear_attention",
      "full_attention",
      "linear_attention",
      "linear_attention",
      "linear_attention",
      "full_attention",
      "linear_attention",
      "linear_attention",
      "linear_attention",
      "full_attention",
      "linear_attention",
      "linear_attention",
      "linear_attention",
      "full_attention",
      "linear_attention",
      "linear_attention",
      "linear_attention",
      "full_attention",
      "linear_attention",
      "linear_attention",
      "linear_attention",
      "full_attention",
      "linear_attention",
      "linear_attention",
      "linear_attention",
      "full_attention"
    ],
    "linear_conv_kernel_dim": 4,
    "linear_key_head_dim": 128,
    "linear_num_key_heads": 16,
    "linear_num_value_heads": 32,
    "linear_value_head_dim": 128,
    "mamba_ssm_dtype": "float32",
    "max_position_embeddings": 262144,
    "mlp_only_layers": [],
    "model_type": "qwen3_5_text",
    "mtp_num_hidden_layers": 1,
    "mtp_use_dedicated_embeddings": false,
    "num_attention_heads": 16,
    "num_hidden_layers": 32,
    "num_key_value_heads": 4,
    "pad_token_id": null,
    "partial_rotary_factor": 0.25,
    "rms_norm_eps": 1e-06,
    "rope_parameters": {
      "mrope_interleaved": true,
      "mrope_section": [
        11,
        11,
        10
      ],
      "partial_rotary_factor": 0.25,
      "rope_theta": 10000000,
      "rope_type": "default"
    },
    "tie_word_embeddings": true,
    "use_cache": true,
    "vocab_size": 248320
  },
  "tie_word_embeddings": true,
  "transformers_version": "5.3.0",
  "use_cache": false,
  "use_geometry_encoder": true,
  "video_token_id": 248057,
  "vision_config": {
    "deepstack_visual_indexes": [],
    "depth": 24,
    "dtype": "bfloat16",
    "hidden_act": "gelu_pytorch_tanh",
    "hidden_size": 1024,
    "in_channels": 3,
    "initializer_range": 0.02,
    "intermediate_size": 4096,
    "model_type": "qwen3_5",
    "num_heads": 16,
    "num_position_embeddings": 2304,
    "out_hidden_size": 2560,
    "patch_size": 16,
    "spatial_merge_size": 2,
    "temporal_patch_size": 2
  },
  "vision_end_token_id": 248054,
  "vision_language_fusion_layers": null,
  "vision_start_token_id": 248053
}

Loading datasets: [{'annotation_path': '/mnt/data/vmo-ai-task/anhdh35/JanusVLN/train_r2r_rxr.json', 'data_path': '/mnt/data/vmo-ai-task/anhdh35/JanusVLN', 'tag': '3d', 'sampling_rate': 1.0, 'dataset_name': 'train_r2r_rxr'}]
Total training samples: 2429620
Total training samples: 2429620
Total training samples: 2429620
Total training samples: 2429620
Formatting inputs...Skip in lazy mode
Formatting inputs...Skip in lazy mode
Formatting inputs...Skip in lazy mode
Formatting inputs...Skip in lazy mode
The tokenizer has new PAD/BOS/EOS tokens that differ from the model config and generation config. The model config and generation config were aligned accordingly, being updated with the tokenizer's values. Updated tokens: {'eos_token_id': 248046, 'pad_token_id': 248044}.
The tokenizer has new PAD/BOS/EOS tokens that differ from the model config and generation config. The model config and generation config were aligned accordingly, being updated with the tokenizer's values. Updated tokens: {'eos_token_id': 248046, 'pad_token_id': 248044}.
The tokenizer has new PAD/BOS/EOS tokens that differ from the model config and generation config. The model config and generation config were aligned accordingly, being updated with the tokenizer's values. Updated tokens: {'eos_token_id': 248046, 'pad_token_id': 248044}.
The tokenizer has new PAD/BOS/EOS tokens that differ from the model config and generation config. The model config and generation config were aligned accordingly, being updated with the tokenizer's values. Updated tokens: {'eos_token_id': 248046, 'pad_token_id': 248044}.
NCCL version 2.26.2+cuda12.2
[GeometryBatch] micro_batch=4, mode=per-sample VGGT fallback, input_shapes=[(9, 3, 420, 560), (9, 3, 420, 560), (9, 3, 420, 560), (4, 3, 420, 560)]
/home/dungpq6/Project/SpatialStack/src/qwen_vl/model/geometry_encoders/vggt_encoder.py:337: FutureWarning: `torch.cuda.amp.autocast(args...)` is deprecated. Please use `torch.amp.autocast('cuda', args...)` instead.
  with torch.cuda.amp.autocast(dtype=dtype):

  0%|          | 0/30 [00:00<?, ?it/s][GeometryBatch] micro_batch=4, mode=per-sample VGGT fallback, input_shapes=[(9, 3, 420, 560), (9, 3, 420, 560), (9, 3, 420, 560), (3, 3, 420, 560)]
/home/dungpq6/Project/SpatialStack/src/qwen_vl/model/geometry_encoders/vggt_encoder.py:337: FutureWarning: `torch.cuda.amp.autocast(args...)` is deprecated. Please use `torch.amp.autocast('cuda', args...)` instead.
  with torch.cuda.amp.autocast(dtype=dtype):
[GeometryBatch] micro_batch=4, mode=per-sample VGGT fallback, input_shapes=[(9, 3, 420, 560), (9, 3, 420, 560), (9, 3, 420, 560), (3, 3, 420, 560)]
/home/dungpq6/Project/SpatialStack/src/qwen_vl/model/geometry_encoders/vggt_encoder.py:337: FutureWarning: `torch.cuda.amp.autocast(args...)` is deprecated. Please use `torch.amp.autocast('cuda', args...)` instead.
  with torch.cuda.amp.autocast(dtype=dtype):
[GeometryBatch] micro_batch=4, mode=per-sample VGGT fallback, input_shapes=[(9, 3, 420, 560), (9, 3, 420, 560), (9, 3, 420, 560), (2, 3, 420, 560)]
/home/dungpq6/Project/SpatialStack/src/qwen_vl/model/geometry_encoders/vggt_encoder.py:337: FutureWarning: `torch.cuda.amp.autocast(args...)` is deprecated. Please use `torch.amp.autocast('cuda', args...)` instead.
  with torch.cuda.amp.autocast(dtype=dtype):
[GeometryBatch] micro_batch=4, mode=batched VGGT, input_shapes=[(9, 3, 420, 560), (9, 3, 420, 560), (9, 3, 420, 560), (9, 3, 420, 560)]
/home/dungpq6/Project/SpatialStack/src/qwen_vl/model/geometry_encoders/vggt_encoder.py:480: FutureWarning: `torch.cuda.amp.autocast(args...)` is deprecated. Please use `torch.amp.autocast('cuda', args...)` instead.
  with torch.cuda.amp.autocast(dtype=dtype):
[GeometryBatch] micro_batch=4, mode=batched VGGT, input_shapes=[(9, 3, 420, 560), (9, 3, 420, 560), (9, 3, 420, 560), (9, 3, 420, 560)]
/home/dungpq6/Project/SpatialStack/src/qwen_vl/model/geometry_encoders/vggt_encoder.py:480: FutureWarning: `torch.cuda.amp.autocast(args...)` is deprecated. Please use `torch.amp.autocast('cuda', args...)` instead.
  with torch.cuda.amp.autocast(dtype=dtype):
[GeometryBatch] micro_batch=4, mode=batched VGGT, input_shapes=[(9, 3, 420, 560), (9, 3, 420, 560), (9, 3, 420, 560), (9, 3, 420, 560)]
/home/dungpq6/Project/SpatialStack/src/qwen_vl/model/geometry_encoders/vggt_encoder.py:480: FutureWarning: `torch.cuda.amp.autocast(args...)` is deprecated. Please use `torch.amp.autocast('cuda', args...)` instead.
  with torch.cuda.amp.autocast(dtype=dtype):
[GeometryBatch] micro_batch=4, mode=batched VGGT, input_shapes=[(9, 3, 420, 560), (9, 3, 420, 560), (9, 3, 420, 560), (9, 3, 420, 560)]
/home/dungpq6/Project/SpatialStack/src/qwen_vl/model/geometry_encoders/vggt_encoder.py:480: FutureWarning: `torch.cuda.amp.autocast(args...)` is deprecated. Please use `torch.amp.autocast('cuda', args...)` instead.
  with torch.cuda.amp.autocast(dtype=dtype):

  3%|▎         | 1/30 [00:50<24:11, 50.06s/it]
  7%|▋         | 2/30 [01:04<13:31, 28.99s/it]
 10%|█         | 3/30 [01:19<10:11, 22.65s/it]
 13%|█▎        | 4/30 [01:43<10:02, 23.17s/it]
 17%|█▋        | 5/30 [01:57<08:19, 19.98s/it]
 20%|██        | 6/30 [02:12<07:16, 18.19s/it]
 23%|██▎       | 7/30 [02:26<06:30, 16.98s/it]
 27%|██▋       | 8/30 [02:42<06:04, 16.56s/it]
 30%|███       | 9/30 [02:56<05:31, 15.77s/it]
 33%|███▎      | 10/30 [03:12<05:13, 15.66s/it]
                                               
{'loss': '0.9592', 'grad_norm': '6.966', 'learning_rate': '1.647e-05', 'epoch': '0.0002634'}

 33%|███▎      | 10/30 [03:12<05:13, 15.66s/it]
 37%|███▋      | 11/30 [03:27<04:55, 15.55s/it][TrainingProfile] step=11 wall_s=14.926 file=/mnt/data/vmo-ai-task/dungpq6/model-checkpoint/spatialstack_janus_VGGT-Optimized_test/output/batch_profile.BDZIPD/run4_bs4/profile_rank0.jsonl

 40%|████      | 12/30 [03:42<04:37, 15.42s/it][TrainingProfile] step=12 wall_s=15.125 file=/mnt/data/vmo-ai-task/dungpq6/model-checkpoint/spatialstack_janus_VGGT-Optimized_test/output/batch_profile.BDZIPD/run4_bs4/profile_rank0.jsonl

 43%|████▎     | 13/30 [03:56<04:17, 15.12s/it][TrainingProfile] step=13 wall_s=14.431 file=/mnt/data/vmo-ai-task/dungpq6/model-checkpoint/spatialstack_janus_VGGT-Optimized_test/output/batch_profile.BDZIPD/run4_bs4/profile_rank0.jsonl

 47%|████▋     | 14/30 [04:10<03:56, 14.80s/it][TrainingProfile] step=14 wall_s=14.038 file=/mnt/data/vmo-ai-task/dungpq6/model-checkpoint/spatialstack_janus_VGGT-Optimized_test/output/batch_profile.BDZIPD/run4_bs4/profile_rank0.jsonl

 50%|█████     | 15/30 [04:25<03:41, 14.74s/it][TrainingProfile] step=15 wall_s=14.606 file=/mnt/data/vmo-ai-task/dungpq6/model-checkpoint/spatialstack_janus_VGGT-Optimized_test/output/batch_profile.BDZIPD/run4_bs4/profile_rank0.jsonl

 53%|█████▎    | 16/30 [04:40<03:25, 14.66s/it][TrainingProfile] step=16 wall_s=14.483 file=/mnt/data/vmo-ai-task/dungpq6/model-checkpoint/spatialstack_janus_VGGT-Optimized_test/output/batch_profile.BDZIPD/run4_bs4/profile_rank0.jsonl

 57%|█████▋    | 17/30 [04:55<03:13, 14.91s/it][TrainingProfile] step=17 wall_s=15.496 file=/mnt/data/vmo-ai-task/dungpq6/model-checkpoint/spatialstack_janus_VGGT-Optimized_test/output/batch_profile.BDZIPD/run4_bs4/profile_rank0.jsonl

 60%|██████    | 18/30 [05:10<02:58, 14.90s/it][TrainingProfile] step=18 wall_s=14.875 file=/mnt/data/vmo-ai-task/dungpq6/model-checkpoint/spatialstack_janus_VGGT-Optimized_test/output/batch_profile.BDZIPD/run4_bs4/profile_rank0.jsonl

 63%|██████▎   | 19/30 [05:24<02:42, 14.81s/it][TrainingProfile] step=19 wall_s=14.581 file=/mnt/data/vmo-ai-task/dungpq6/model-checkpoint/spatialstack_janus_VGGT-Optimized_test/output/batch_profile.BDZIPD/run4_bs4/profile_rank0.jsonl

 67%|██████▋   | 20/30 [05:40<02:29, 14.97s/it][TrainingProfile] step=20 wall_s=15.355 file=/mnt/data/vmo-ai-task/dungpq6/model-checkpoint/spatialstack_janus_VGGT-Optimized_test/output/batch_profile.BDZIPD/run4_bs4/profile_rank0.jsonl

                                               
{'loss': '0.2889', 'grad_norm': '4.721', 'learning_rate': '6.299e-06', 'epoch': '0.0005268'}

 67%|██████▋   | 20/30 [05:40<02:29, 14.97s/it]
 70%|███████   | 21/30 [05:55<02:14, 14.95s/it][TrainingProfile] step=21 wall_s=14.889 file=/mnt/data/vmo-ai-task/dungpq6/model-checkpoint/spatialstack_janus_VGGT-Optimized_test/output/batch_profile.BDZIPD/run4_bs4/profile_rank0.jsonl

 73%|███████▎  | 22/30 [06:09<01:57, 14.69s/it][TrainingProfile] step=22 wall_s=14.104 file=/mnt/data/vmo-ai-task/dungpq6/model-checkpoint/spatialstack_janus_VGGT-Optimized_test/output/batch_profile.BDZIPD/run4_bs4/profile_rank0.jsonl

 77%|███████▋  | 23/30 [06:23<01:40, 14.42s/it][TrainingProfile] step=23 wall_s=13.770 file=/mnt/data/vmo-ai-task/dungpq6/model-checkpoint/spatialstack_janus_VGGT-Optimized_test/output/batch_profile.BDZIPD/run4_bs4/profile_rank0.jsonl

 80%|████████  | 24/30 [06:38<01:28, 14.78s/it][TrainingProfile] step=24 wall_s=15.622 file=/mnt/data/vmo-ai-task/dungpq6/model-checkpoint/spatialstack_janus_VGGT-Optimized_test/output/batch_profile.BDZIPD/run4_bs4/profile_rank0.jsonl

 83%|████████▎ | 25/30 [06:52<01:12, 14.51s/it][TrainingProfile] step=25 wall_s=13.885 file=/mnt/data/vmo-ai-task/dungpq6/model-checkpoint/spatialstack_janus_VGGT-Optimized_test/output/batch_profile.BDZIPD/run4_bs4/profile_rank0.jsonl

 87%|████████▋ | 26/30 [07:07<00:58, 14.57s/it][TrainingProfile] step=26 wall_s=14.717 file=/mnt/data/vmo-ai-task/dungpq6/model-checkpoint/spatialstack_janus_VGGT-Optimized_test/output/batch_profile.BDZIPD/run4_bs4/profile_rank0.jsonl

 90%|█████████ | 27/30 [07:21<00:43, 14.53s/it][TrainingProfile] step=27 wall_s=14.423 file=/mnt/data/vmo-ai-task/dungpq6/model-checkpoint/spatialstack_janus_VGGT-Optimized_test/output/batch_profile.BDZIPD/run4_bs4/profile_rank0.jsonl

 93%|█████████▎| 28/30 [07:36<00:29, 14.56s/it][TrainingProfile] step=28 wall_s=14.617 file=/mnt/data/vmo-ai-task/dungpq6/model-checkpoint/spatialstack_janus_VGGT-Optimized_test/output/batch_profile.BDZIPD/run4_bs4/profile_rank0.jsonl

 97%|█████████▋| 29/30 [07:51<00:14, 14.69s/it][TrainingProfile] step=29 wall_s=15.002 file=/mnt/data/vmo-ai-task/dungpq6/model-checkpoint/spatialstack_janus_VGGT-Optimized_test/output/batch_profile.BDZIPD/run4_bs4/profile_rank0.jsonl

100%|██████████| 30/30 [08:06<00:00, 14.71s/it][TrainingProfile] step=30 wall_s=14.755 file=/mnt/data/vmo-ai-task/dungpq6/model-checkpoint/spatialstack_janus_VGGT-Optimized_test/output/batch_profile.BDZIPD/run4_bs4/profile_rank0.jsonl

                                               
{'loss': '0.276', 'grad_norm': '7.161', 'learning_rate': '5.862e-08', 'epoch': '0.0007902'}

100%|██████████| 30/30 [08:06<00:00, 14.71s/it]

Writing model shards:   0%|          | 0/1 [00:00<?, ?it/s][A

Writing model shards: 100%|██████████| 1/1 [00:10<00:00, 10.71s/it][A
Writing model shards: 100%|██████████| 1/1 [00:10<00:00, 10.72s/it]

                                               
{'train_runtime': '543.9', 'train_samples_per_second': '3.53', 'train_steps_per_second': '0.055', 'train_loss': '0.508', 'epoch': '0.0007902'}

100%|██████████| 30/30 [09:03<00:00, 14.71s/it]
100%|██████████| 30/30 [09:03<00:00, 18.13s/it]

Writing model shards:   0%|          | 0/1 [00:00<?, ?it/s]
Writing model shards: 100%|██████████| 1/1 [00:09<00:00,  9.21s/it]
Writing model shards: 100%|██████████| 1/1 [00:09<00:00,  9.21s/it]
[rank0]:[W918 05:29:58.693105872 ProcessGroupNCCL.cpp:1476] Warning: WARNING: destroy_process_group() was not called before program exit, which can leak resources. For more info, please see https://pytorch.org/docs/stable/distributed.html#shutdown (function operator())