Hugging Face's logo

Datasets:
phamquandung
/
debugging_optimized

Dataset card Files
xet
 Community
phamquandung commited on
Commit
04a5b46
·
verified ·
1 Parent(s): 40dafeb

Upload folder using huggingface_hub

Browse files
Files changed (1) hide show
  1. bs16_low_prefetch/train.log +731 -0
bs16_low_prefetch/train.log ADDED
@@ -0,0 +1,731 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
0
  0%| | 0/12 [00:00<?, ?it/s][GeometryBatch] micro_batch=16, mode=per-sample VGGT fallback, input_shapes=[(9, 3, 420, 560), (9, 3, 420, 560), (9, 3, 420, 560), (9, 3, 420, 560), (9, 3, 420, 560), (9, 3, 420, 560), (9, 3, 420, 560), (9, 3, 420, 560), (9, 3, 420, 560), (9, 3, 420, 560), (9, 3, 420, 560), (9, 3, 420, 560), (9, 3, 420, 560), (9, 3, 420, 560), (6, 3, 420, 560), (3, 3, 420, 560)]
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
  8%|▊ | 1/12 [01:09<12:49, 69.94s/it][TrainingMemory] {"rank": "1", "stage": "before_logging", "step": 1, "allocated_GiB": 24.172, "reserved_GiB": 71.883, "peak_allocated_GiB": 46.028, "peak_reserved_GiB": 71.883, "device_free_GiB": 1.042, "device_total_GiB": 79.189}
 
 
 
 
 
 
 
 
 
 
 
 
 
 
2
  17%|█▋ | 2/12 [01:49<08:40, 52.06s/it]/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.
 
 
 
 
 
3
  25%|██▌ | 3/12 [02:13<05:54, 39.36s/it][TrainingMemory] {"rank": "3", "stage": "before_logging", "step": 3, "allocated_GiB": 24.21, "reserved_GiB": 63.305, "peak_allocated_GiB": 54.124, "peak_reserved_GiB": 63.305, "device_free_GiB": 9.62, "device_total_GiB": 79.189}
 
 
 
 
 
4
  33%|███▎ | 4/12 [02:37<04:26, 33.30s/it][TrainingMemory] {"rank": "3", "stage": "before_logging", "step": 5, "allocated_GiB": 24.182, "reserved_GiB": 65.004, "peak_allocated_GiB": 54.512, "peak_reserved_GiB": 65.004, "device_free_GiB": 7.919, "device_total_GiB": 79.189}
 
 
 
 
5
  42%|████▏ | 5/12 [02:53<03:08, 26.86s/it][TrainingMemory] {"rank": "2", "stage": "before_logging", "step": 6, "allocated_GiB": 24.182, "reserved_GiB": 71.924, "peak_allocated_GiB": 54.164, "peak_reserved_GiB": 71.924, "device_free_GiB": 0.999, "device_total_GiB": 79.189}
 
 
6
  50%|█████ | 6/12 [03:08<02:17, 22.90s/it][TrainingMemory] {"rank": "1", "stage": "before_logging", "step": 6, "allocated_GiB": 24.201, "reserved_GiB": 71.883, "peak_allocated_GiB": 54.416, "peak_reserved_GiB": 71.883, "device_free_GiB": 1.04, "device_total_GiB": 79.189}
 
 
 
 
 
 
 
 
 
7
  58%|█████▊ | 7/12 [03:23<01:42, 20.45s/it][TrainingMemory] {"rank": "1", "stage": "before_logging", "step": 7, "allocated_GiB": 24.201, "reserved_GiB": 71.883, "peak_allocated_GiB": 54.855, "peak_reserved_GiB": 71.883, "device_free_GiB": 1.038, "device_total_GiB": 79.189}
 
 
 
8
  67%|██████▋ | 8/12 [04:03<01:46, 26.65s/it][TrainingMemory] {"rank": "3", "stage": "before_logging", "step": 8, "allocated_GiB": 24.182, "reserved_GiB": 65.004, "peak_allocated_GiB": 54.43, "peak_reserved_GiB": 65.004, "device_free_GiB": 7.919, "device_total_GiB": 79.189}
 
 
 
 
 
 
 
9
  75%|███████▌ | 9/12 [04:19<01:09, 23.24s/it][TrainingMemory] {"rank": "3", "stage": "before_logging", "step": 9, "allocated_GiB": 24.249, "reserved_GiB": 65.848, "peak_allocated_GiB": 54.738, "peak_reserved_GiB": 65.848, "device_free_GiB": 7.075, "device_total_GiB": 79.189}
 
 
 
 
 
 
 
10
  83%|████████▎ | 10/12 [04:34<00:41, 20.85s/it][TrainingMemory] {"rank": "1", "stage": "after_logging", "step": 10, "allocated_GiB": 24.211, "reserved_GiB": 71.883, "peak_allocated_GiB": 54.644, "peak_reserved_GiB": 71.883, "device_free_GiB": 0.565, "device_total_GiB": 79.189}
 
 
 
11
 
 
 
12
  83%|████████▎ | 10/12 [04:35<00:41, 20.85s/it][TrainingMemory] {"rank": "2", "stage": "before_logging", "step": 11, "allocated_GiB": 24.144, "reserved_GiB": 71.924, "peak_allocated_GiB": 54.433, "peak_reserved_GiB": 71.924, "device_free_GiB": 0.524, "device_total_GiB": 79.189}
 
 
 
13
  92%|█████████▏| 11/12 [04:50<00:19, 19.24s/it][TrainingMemory] {"rank": "3", "stage": "before_logging", "step": 11, "allocated_GiB": 24.154, "reserved_GiB": 65.848, "peak_allocated_GiB": 54.197, "peak_reserved_GiB": 65.848, "device_free_GiB": 6.602, "device_total_GiB": 79.189}
 
 
 
 
 
 
 
 
 
 
 
 
14
 
 
 
 
 
1
+ [INFO] Running in WANDB offline mode
2
+ [INFO] Running in WANDB offline mode
3
+ [INFO] Running in WANDB offline mode
4
+ [INFO] Running in WANDB offline mode
5
+ Trainer._get_train_sampler replaced with custom implementation.
6
+ warmup_ratio is deprecated and will be removed in v5.2. Use `warmup_steps` instead.
7
+ Trainer._get_train_sampler replaced with custom implementation.
8
+ Trainer._get_train_sampler replaced with custom implementation.
9
+ Trainer._get_train_sampler replaced with custom implementation.
10
+ warmup_ratio is deprecated and will be removed in v5.2. Use `warmup_steps` instead.
11
+ warmup_ratio is deprecated and will be removed in v5.2. Use `warmup_steps` instead.
12
+ warmup_ratio is deprecated and will be removed in v5.2. Use `warmup_steps` instead.
13
+ [2026-09-18 03:49:17,786] [INFO] [real_accelerator.py:222:get_accelerator] Setting ds_accelerator to cuda (auto detect)
14
+ [2026-09-18 03:49:17,798] [INFO] [real_accelerator.py:222:get_accelerator] Setting ds_accelerator to cuda (auto detect)
15
+ [2026-09-18 03:49:17,799] [INFO] [real_accelerator.py:222:get_accelerator] Setting ds_accelerator to cuda (auto detect)
16
+ [2026-09-18 03:49:17,801] [INFO] [real_accelerator.py:222:get_accelerator] Setting ds_accelerator to cuda (auto detect)
17
+ [2026-09-18 03:49:18,104] [INFO] [comm.py:658:init_distributed] cdb=None
18
+ [2026-09-18 03:49:18,104] [INFO] [comm.py:689:init_distributed] Initializing TorchBackend in DeepSpeed with backend nccl
19
+ [2026-09-18 03:49:18,129] [INFO] [comm.py:658:init_distributed] cdb=None
20
+ [2026-09-18 03:49:18,174] [INFO] [comm.py:658:init_distributed] cdb=None
21
+ [2026-09-18 03:49:18,178] [INFO] [comm.py:658:init_distributed] cdb=None
22
+ [TrainingFix] Qwen3.5 FlashAttention MRoPE packing guard enabled
23
+ [TrainingFix] Qwen3.5 FlashAttention MRoPE packing guard enabled
24
+ [TrainingFix] Qwen3.5 FlashAttention MRoPE packing guard enabled
25
+ [TrainingFix] Qwen3.5 FlashAttention MRoPE packing guard enabled
26
+
27
+
28
+
29
+
30
+ [VGGTEncoder] eval KV-cache window: start=8 recent=48 (total=56 frames)
31
+ [FeatureFusion] geometry fusion delta scaled by 0.5
32
+ [FeatureFusion] importance gate ENABLED on geometry delta
33
+ [FeatureFusion] learnable geometry scale ENABLED (init 0.5)
34
+ [VGGTEncoder] eval KV-cache window: start=8 recent=48 (total=56 frames)
35
+ [FeatureFusion] geometry fusion delta scaled by 0.5
36
+ [FeatureFusion] importance gate ENABLED on geometry delta
37
+ [FeatureFusion] learnable geometry scale ENABLED (init 0.5)
38
+ [VGGTEncoder] eval KV-cache window: start=8 recent=48 (total=56 frames)
39
+ [FeatureFusion] geometry fusion delta scaled by 0.5
40
+ [FeatureFusion] importance gate ENABLED on geometry delta
41
+ [FeatureFusion] learnable geometry scale ENABLED (init 0.5)
42
+ [VGGTEncoder] eval KV-cache window: start=8 recent=48 (total=56 frames)
43
+ [FeatureFusion] geometry fusion delta scaled by 0.5
44
+ [FeatureFusion] importance gate ENABLED on geometry delta
45
+ [FeatureFusion] learnable geometry scale ENABLED (init 0.5)
46
+ Loading weights from local directory
47
+ Loading weights from local directory
48
+ Loading weights from local directory
49
+ Loading weights from local directory
50
+ >>>>> STOP loss up-weight: 3.0 on token ids [48964]
51
+ Qwen3_5Config {
52
+ "architectures": [
53
+ "Qwen3_5ForConditionalGeneration"
54
+ ],
55
+ "dtype": "bfloat16",
56
+ "feature_fusion_method": "deepstack_language_add",
57
+ "fusion_num_layers": 1,
58
+ "geometry_encoder_layers": [
59
+ 11,
60
+ 17,
61
+ 23
62
+ ],
63
+ "geometry_encoder_path": "/mnt/data/vmo-ai-task/dungpq6/model-checkpoint/VGGT-1B",
64
+ "geometry_encoder_streaming": true,
65
+ "geometry_encoder_type": "vggt",
66
+ "geometry_frame_strict": true,
67
+ "geometry_fusion_layers": [
68
+ 0,
69
+ 1,
70
+ 2
71
+ ],
72
+ "geometry_fusion_scale": 0.5,
73
+ "geometry_importance_gate": true,
74
+ "geometry_learnable_scale": true,
75
+ "geometry_merger_type": "mlp",
76
+ "geometry_spatial_bias": false,
77
+ "image_token_id": 248056,
78
+ "include_camera_token": false,
79
+ "model_type": "qwen3_5",
80
+ "pos_encoding_type": "none",
81
+ "reference_frame": "first",
82
+ "supervised_token_loss": true,
83
+ "text_config": {
84
+ "attention_bias": false,
85
+ "attention_dropout": 0.0,
86
+ "attn_output_gate": true,
87
+ "bos_token_id": null,
88
+ "dtype": "bfloat16",
89
+ "eos_token_id": 248044,
90
+ "full_attention_interval": 4,
91
+ "head_dim": 256,
92
+ "hidden_act": "silu",
93
+ "hidden_size": 2560,
94
+ "initializer_range": 0.02,
95
+ "intermediate_size": 9216,
96
+ "layer_types": [
97
+ "linear_attention",
98
+ "linear_attention",
99
+ "linear_attention",
100
+ "full_attention",
101
+ "linear_attention",
102
+ "linear_attention",
103
+ "linear_attention",
104
+ "full_attention",
105
+ "linear_attention",
106
+ "linear_attention",
107
+ "linear_attention",
108
+ "full_attention",
109
+ "linear_attention",
110
+ "linear_attention",
111
+ "linear_attention",
112
+ "full_attention",
113
+ "linear_attention",
114
+ "linear_attention",
115
+ "linear_attention",
116
+ "full_attention",
117
+ "linear_attention",
118
+ "linear_attention",
119
+ "linear_attention",
120
+ "full_attention",
121
+ "linear_attention",
122
+ "linear_attention",
123
+ "linear_attention",
124
+ "full_attention",
125
+ "linear_attention",
126
+ "linear_attention",
127
+ "linear_attention",
128
+ "full_attention"
129
+ ],
130
+ "linear_conv_kernel_dim": 4,
131
+ "linear_key_head_dim": 128,
132
+ "linear_num_key_heads": 16,
133
+ "linear_num_value_heads": 32,
134
+ "linear_value_head_dim": 128,
135
+ "mamba_ssm_dtype": "float32",
136
+ "max_position_embeddings": 262144,
137
+ "mlp_only_layers": [],
138
+ "model_type": "qwen3_5_text",
139
+ "mtp_num_hidden_layers": 1,
140
+ "mtp_use_dedicated_embeddings": false,
141
+ "num_attention_heads": 16,
142
+ "num_hidden_layers": 32,
143
+ "num_key_value_heads": 4,
144
+ "pad_token_id": null,
145
+ "partial_rotary_factor": 0.25,
146
+ "rms_norm_eps": 1e-06,
147
+ "rope_parameters": {
148
+ "mrope_interleaved": true,
149
+ "mrope_section": [
150
+ 11,
151
+ 11,
152
+ 10
153
+ ],
154
+ "partial_rotary_factor": 0.25,
155
+ "rope_theta": 10000000,
156
+ "rope_type": "default"
157
+ },
158
+ "tie_word_embeddings": true,
159
+ "use_cache": true,
160
+ "vocab_size": 248320
161
+ },
162
+ "tie_word_embeddings": true,
163
+ "transformers_version": "5.3.0",
164
+ "use_cache": false,
165
+ "use_geometry_encoder": true,
166
+ "video_token_id": 248057,
167
+ "vision_config": {
168
+ "deepstack_visual_indexes": [],
169
+ "depth": 24,
170
+ "dtype": "bfloat16",
171
+ "hidden_act": "gelu_pytorch_tanh",
172
+ "hidden_size": 1024,
173
+ "in_channels": 3,
174
+ "initializer_range": 0.02,
175
+ "intermediate_size": 4096,
176
+ "model_type": "qwen3_5",
177
+ "num_heads": 16,
178
+ "num_position_embeddings": 2304,
179
+ "out_hidden_size": 2560,
180
+ "patch_size": 16,
181
+ "spatial_merge_size": 2,
182
+ "temporal_patch_size": 2
183
+ },
184
+ "vision_end_token_id": 248054,
185
+ "vision_language_fusion_layers": null,
186
+ "vision_start_token_id": 248053
187
+ }
188
+
189
+ 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'}]
190
+ >>>>> STOP loss up-weight: 3.0 on token ids [48964]
191
+ Qwen3_5Config {
192
+ "architectures": [
193
+ "Qwen3_5ForConditionalGeneration"
194
+ ],
195
+ "dtype": "bfloat16",
196
+ "feature_fusion_method": "deepstack_language_add",
197
+ "fusion_num_layers": 1,
198
+ "geometry_encoder_layers": [
199
+ 11,
200
+ 17,
201
+ 23
202
+ ],
203
+ "geometry_encoder_path": "/mnt/data/vmo-ai-task/dungpq6/model-checkpoint/VGGT-1B",
204
+ "geometry_encoder_streaming": true,
205
+ "geometry_encoder_type": "vggt",
206
+ "geometry_frame_strict": true,
207
+ "geometry_fusion_layers": [
208
+ 0,
209
+ 1,
210
+ 2
211
+ ],
212
+ "geometry_fusion_scale": 0.5,
213
+ "geometry_importance_gate": true,
214
+ "geometry_learnable_scale": true,
215
+ "geometry_merger_type": "mlp",
216
+ "geometry_spatial_bias": false,
217
+ "image_token_id": 248056,
218
+ "include_camera_token": false,
219
+ "model_type": "qwen3_5",
220
+ "pos_encoding_type": "none",
221
+ "reference_frame": "first",
222
+ "supervised_token_loss": true,
223
+ "text_config": {
224
+ "attention_bias": false,
225
+ "attention_dropout": 0.0,
226
+ "attn_output_gate": true,
227
+ "bos_token_id": null,
228
+ "dtype": "bfloat16",
229
+ "eos_token_id": 248044,
230
+ "full_attention_interval": 4,
231
+ "head_dim": 256,
232
+ "hidden_act": "silu",
233
+ "hidden_size": 2560,
234
+ "initializer_range": 0.02,
235
+ "intermediate_size": 9216,
236
+ "layer_types": [
237
+ "linear_attention",
238
+ "linear_attention",
239
+ "linear_attention",
240
+ "full_attention",
241
+ "linear_attention",
242
+ "linear_attention",
243
+ "linear_attention",
244
+ "full_attention",
245
+ "linear_attention",
246
+ "linear_attention",
247
+ "linear_attention",
248
+ "full_attention",
249
+ "linear_attention",
250
+ "linear_attention",
251
+ "linear_attention",
252
+ "full_attention",
253
+ "linear_attention",
254
+ "linear_attention",
255
+ "linear_attention",
256
+ "full_attention",
257
+ "linear_attention",
258
+ "linear_attention",
259
+ "linear_attention",
260
+ "full_attention",
261
+ "linear_attention",
262
+ "linear_attention",
263
+ "linear_attention",
264
+ "full_attention",
265
+ "linear_attention",
266
+ "linear_attention",
267
+ "linear_attention",
268
+ "full_attention"
269
+ ],
270
+ "linear_conv_kernel_dim": 4,
271
+ "linear_key_head_dim": 128,
272
+ "linear_num_key_heads": 16,
273
+ "linear_num_value_heads": 32,
274
+ "linear_value_head_dim": 128,
275
+ "mamba_ssm_dtype": "float32",
276
+ "max_position_embeddings": 262144,
277
+ "mlp_only_layers": [],
278
+ "model_type": "qwen3_5_text",
279
+ "mtp_num_hidden_layers": 1,
280
+ "mtp_use_dedicated_embeddings": false,
281
+ "num_attention_heads": 16,
282
+ "num_hidden_layers": 32,
283
+ "num_key_value_heads": 4,
284
+ "pad_token_id": null,
285
+ "partial_rotary_factor": 0.25,
286
+ "rms_norm_eps": 1e-06,
287
+ "rope_parameters": {
288
+ "mrope_interleaved": true,
289
+ "mrope_section": [
290
+ 11,
291
+ 11,
292
+ 10
293
+ ],
294
+ "partial_rotary_factor": 0.25,
295
+ "rope_theta": 10000000,
296
+ "rope_type": "default"
297
+ },
298
+ "tie_word_embeddings": true,
299
+ "use_cache": true,
300
+ "vocab_size": 248320
301
+ },
302
+ "tie_word_embeddings": true,
303
+ "transformers_version": "5.3.0",
304
+ "use_cache": false,
305
+ "use_geometry_encoder": true,
306
+ "video_token_id": 248057,
307
+ "vision_config": {
308
+ "deepstack_visual_indexes": [],
309
+ "depth": 24,
310
+ "dtype": "bfloat16",
311
+ "hidden_act": "gelu_pytorch_tanh",
312
+ "hidden_size": 1024,
313
+ "in_channels": 3,
314
+ "initializer_range": 0.02,
315
+ "intermediate_size": 4096,
316
+ "model_type": "qwen3_5",
317
+ "num_heads": 16,
318
+ "num_position_embeddings": 2304,
319
+ "out_hidden_size": 2560,
320
+ "patch_size": 16,
321
+ "spatial_merge_size": 2,
322
+ "temporal_patch_size": 2
323
+ },
324
+ "vision_end_token_id": 248054,
325
+ "vision_language_fusion_layers": null,
326
+ "vision_start_token_id": 248053
327
+ }
328
+
329
+ 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'}]
330
+ >>>>> STOP loss up-weight: 3.0 on token ids [48964]
331
+ Vision Module - Attention Blocks:
332
+ Trainable Block Indices: None
333
+ 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]
334
+ Merger Module Trainable: True
335
+ LLM Module - Embed Tokens Trainable: True
336
+ 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]
337
+ LLM Module - Non-Trainable Layer Indices: None
338
+ Qwen3_5Config {
339
+ "architectures": [
340
+ "Qwen3_5ForConditionalGeneration"
341
+ ],
342
+ "dtype": "bfloat16",
343
+ "feature_fusion_method": "deepstack_language_add",
344
+ "fusion_num_layers": 1,
345
+ "geometry_encoder_layers": [
346
+ 11,
347
+ 17,
348
+ 23
349
+ ],
350
+ "geometry_encoder_path": "/mnt/data/vmo-ai-task/dungpq6/model-checkpoint/VGGT-1B",
351
+ "geometry_encoder_streaming": true,
352
+ "geometry_encoder_type": "vggt",
353
+ "geometry_frame_strict": true,
354
+ "geometry_fusion_layers": [
355
+ 0,
356
+ 1,
357
+ 2
358
+ ],
359
+ "geometry_fusion_scale": 0.5,
360
+ "geometry_importance_gate": true,
361
+ "geometry_learnable_scale": true,
362
+ "geometry_merger_type": "mlp",
363
+ "geometry_spatial_bias": false,
364
+ "image_token_id": 248056,
365
+ "include_camera_token": false,
366
+ "model_type": "qwen3_5",
367
+ "pos_encoding_type": "none",
368
+ "reference_frame": "first",
369
+ "supervised_token_loss": true,
370
+ "text_config": {
371
+ "attention_bias": false,
372
+ "attention_dropout": 0.0,
373
+ "attn_output_gate": true,
374
+ "bos_token_id": null,
375
+ "dtype": "bfloat16",
376
+ "eos_token_id": 248044,
377
+ "full_attention_interval": 4,
378
+ "head_dim": 256,
379
+ "hidden_act": "silu",
380
+ "hidden_size": 2560,
381
+ "initializer_range": 0.02,
382
+ "intermediate_size": 9216,
383
+ "layer_types": [
384
+ "linear_attention",
385
+ "linear_attention",
386
+ "linear_attention",
387
+ "full_attention",
388
+ "linear_attention",
389
+ "linear_attention",
390
+ "linear_attention",
391
+ "full_attention",
392
+ "linear_attention",
393
+ "linear_attention",
394
+ "linear_attention",
395
+ "full_attention",
396
+ "linear_attention",
397
+ "linear_attention",
398
+ "linear_attention",
399
+ "full_attention",
400
+ "linear_attention",
401
+ "linear_attention",
402
+ "linear_attention",
403
+ "full_attention",
404
+ "linear_attention",
405
+ "linear_attention",
406
+ "linear_attention",
407
+ "full_attention",
408
+ "linear_attention",
409
+ "linear_attention",
410
+ "linear_attention",
411
+ "full_attention",
412
+ "linear_attention",
413
+ "linear_attention",
414
+ "linear_attention",
415
+ "full_attention"
416
+ ],
417
+ "linear_conv_kernel_dim": 4,
418
+ "linear_key_head_dim": 128,
419
+ "linear_num_key_heads": 16,
420
+ "linear_num_value_heads": 32,
421
+ "linear_value_head_dim": 128,
422
+ "mamba_ssm_dtype": "float32",
423
+ "max_position_embeddings": 262144,
424
+ "mlp_only_layers": [],
425
+ "model_type": "qwen3_5_text",
426
+ "mtp_num_hidden_layers": 1,
427
+ "mtp_use_dedicated_embeddings": false,
428
+ "num_attention_heads": 16,
429
+ "num_hidden_layers": 32,
430
+ "num_key_value_heads": 4,
431
+ "pad_token_id": null,
432
+ "partial_rotary_factor": 0.25,
433
+ "rms_norm_eps": 1e-06,
434
+ "rope_parameters": {
435
+ "mrope_interleaved": true,
436
+ "mrope_section": [
437
+ 11,
438
+ 11,
439
+ 10
440
+ ],
441
+ "partial_rotary_factor": 0.25,
442
+ "rope_theta": 10000000,
443
+ "rope_type": "default"
444
+ },
445
+ "tie_word_embeddings": true,
446
+ "use_cache": true,
447
+ "vocab_size": 248320
448
+ },
449
+ "tie_word_embeddings": true,
450
+ "transformers_version": "5.3.0",
451
+ "use_cache": false,
452
+ "use_geometry_encoder": true,
453
+ "video_token_id": 248057,
454
+ "vision_config": {
455
+ "deepstack_visual_indexes": [],
456
+ "depth": 24,
457
+ "dtype": "bfloat16",
458
+ "hidden_act": "gelu_pytorch_tanh",
459
+ "hidden_size": 1024,
460
+ "in_channels": 3,
461
+ "initializer_range": 0.02,
462
+ "intermediate_size": 4096,
463
+ "model_type": "qwen3_5",
464
+ "num_heads": 16,
465
+ "num_position_embeddings": 2304,
466
+ "out_hidden_size": 2560,
467
+ "patch_size": 16,
468
+ "spatial_merge_size": 2,
469
+ "temporal_patch_size": 2
470
+ },
471
+ "vision_end_token_id": 248054,
472
+ "vision_language_fusion_layers": null,
473
+ "vision_start_token_id": 248053
474
+ }
475
+
476
+ 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'}]
477
+ >>>>> STOP loss up-weight: 3.0 on token ids [48964]
478
+ Qwen3_5Config {
479
+ "architectures": [
480
+ "Qwen3_5ForConditionalGeneration"
481
+ ],
482
+ "dtype": "bfloat16",
483
+ "feature_fusion_method": "deepstack_language_add",
484
+ "fusion_num_layers": 1,
485
+ "geometry_encoder_layers": [
486
+ 11,
487
+ 17,
488
+ 23
489
+ ],
490
+ "geometry_encoder_path": "/mnt/data/vmo-ai-task/dungpq6/model-checkpoint/VGGT-1B",
491
+ "geometry_encoder_streaming": true,
492
+ "geometry_encoder_type": "vggt",
493
+ "geometry_frame_strict": true,
494
+ "geometry_fusion_layers": [
495
+ 0,
496
+ 1,
497
+ 2
498
+ ],
499
+ "geometry_fusion_scale": 0.5,
500
+ "geometry_importance_gate": true,
501
+ "geometry_learnable_scale": true,
502
+ "geometry_merger_type": "mlp",
503
+ "geometry_spatial_bias": false,
504
+ "image_token_id": 248056,
505
+ "include_camera_token": false,
506
+ "model_type": "qwen3_5",
507
+ "pos_encoding_type": "none",
508
+ "reference_frame": "first",
509
+ "supervised_token_loss": true,
510
+ "text_config": {
511
+ "attention_bias": false,
512
+ "attention_dropout": 0.0,
513
+ "attn_output_gate": true,
514
+ "bos_token_id": null,
515
+ "dtype": "bfloat16",
516
+ "eos_token_id": 248044,
517
+ "full_attention_interval": 4,
518
+ "head_dim": 256,
519
+ "hidden_act": "silu",
520
+ "hidden_size": 2560,
521
+ "initializer_range": 0.02,
522
+ "intermediate_size": 9216,
523
+ "layer_types": [
524
+ "linear_attention",
525
+ "linear_attention",
526
+ "linear_attention",
527
+ "full_attention",
528
+ "linear_attention",
529
+ "linear_attention",
530
+ "linear_attention",
531
+ "full_attention",
532
+ "linear_attention",
533
+ "linear_attention",
534
+ "linear_attention",
535
+ "full_attention",
536
+ "linear_attention",
537
+ "linear_attention",
538
+ "linear_attention",
539
+ "full_attention",
540
+ "linear_attention",
541
+ "linear_attention",
542
+ "linear_attention",
543
+ "full_attention",
544
+ "linear_attention",
545
+ "linear_attention",
546
+ "linear_attention",
547
+ "full_attention",
548
+ "linear_attention",
549
+ "linear_attention",
550
+ "linear_attention",
551
+ "full_attention",
552
+ "linear_attention",
553
+ "linear_attention",
554
+ "linear_attention",
555
+ "full_attention"
556
+ ],
557
+ "linear_conv_kernel_dim": 4,
558
+ "linear_key_head_dim": 128,
559
+ "linear_num_key_heads": 16,
560
+ "linear_num_value_heads": 32,
561
+ "linear_value_head_dim": 128,
562
+ "mamba_ssm_dtype": "float32",
563
+ "max_position_embeddings": 262144,
564
+ "mlp_only_layers": [],
565
+ "model_type": "qwen3_5_text",
566
+ "mtp_num_hidden_layers": 1,
567
+ "mtp_use_dedicated_embeddings": false,
568
+ "num_attention_heads": 16,
569
+ "num_hidden_layers": 32,
570
+ "num_key_value_heads": 4,
571
+ "pad_token_id": null,
572
+ "partial_rotary_factor": 0.25,
573
+ "rms_norm_eps": 1e-06,
574
+ "rope_parameters": {
575
+ "mrope_interleaved": true,
576
+ "mrope_section": [
577
+ 11,
578
+ 11,
579
+ 10
580
+ ],
581
+ "partial_rotary_factor": 0.25,
582
+ "rope_theta": 10000000,
583
+ "rope_type": "default"
584
+ },
585
+ "tie_word_embeddings": true,
586
+ "use_cache": true,
587
+ "vocab_size": 248320
588
+ },
589
+ "tie_word_embeddings": true,
590
+ "transformers_version": "5.3.0",
591
+ "use_cache": false,
592
+ "use_geometry_encoder": true,
593
+ "video_token_id": 248057,
594
+ "vision_config": {
595
+ "deepstack_visual_indexes": [],
596
+ "depth": 24,
597
+ "dtype": "bfloat16",
598
+ "hidden_act": "gelu_pytorch_tanh",
599
+ "hidden_size": 1024,
600
+ "in_channels": 3,
601
+ "initializer_range": 0.02,
602
+ "intermediate_size": 4096,
603
+ "model_type": "qwen3_5",
604
+ "num_heads": 16,
605
+ "num_position_embeddings": 2304,
606
+ "out_hidden_size": 2560,
607
+ "patch_size": 16,
608
+ "spatial_merge_size": 2,
609
+ "temporal_patch_size": 2
610
+ },
611
+ "vision_end_token_id": 248054,
612
+ "vision_language_fusion_layers": null,
613
+ "vision_start_token_id": 248053
614
+ }
615
+
616
+ 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'}]
617
+ Total training samples: 2429620
618
+ Total training samples: 2429620
619
+ Total training samples: 2429620
620
+ Total training samples: 2429620
621
+ Formatting inputs...Skip in lazy mode
622
+ Formatting inputs...Skip in lazy mode
623
+ Formatting inputs...Skip in lazy mode
624
+ Formatting inputs...Skip in lazy mode
625
+ 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}.
626
+ 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}.
627
+ 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}.
628
+ 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}.
629
+ NCCL version 2.26.2+cuda12.2
630
+ [TrainingRuntime] {"rank": "1", "versions": {"torch": "2.7.0+cu128", "transformers": "5.3.0", "deepspeed": "0.16.4", "flash-attn": "2.8.3", "triton": "3.3.0", "flash-linear-attention": "0.5.1", "causal-conv1d": "1.6.2.post1"}, "cuda": "12.8", "world_size": 4, "gpu": "NVIDIA H100 80GB HBM3", "micro_batch": 16, "accumulation": 1, "effective_global_batch": 64, "supervised_token_loss": true, "NCCL_NVLS_ENABLE": "0", "CUDA_LAUNCH_BLOCKING": "0", "deepspeed": "scripts/zero2_h100.json"}
631
+ [TrainingMemory] {"rank": "1", "stage": "train_begin", "step": 0, "allocated_GiB": 14.601, "reserved_GiB": 19.906, "peak_allocated_GiB": 18.713, "peak_reserved_GiB": 19.906, "device_free_GiB": 57.864, "device_total_GiB": 79.189}
632
+ [TrainingRuntime] {"rank": "3", "versions": {"torch": "2.7.0+cu128", "transformers": "5.3.0", "deepspeed": "0.16.4", "flash-attn": "2.8.3", "triton": "3.3.0", "flash-linear-attention": "0.5.1", "causal-conv1d": "1.6.2.post1"}, "cuda": "12.8", "world_size": 4, "gpu": "NVIDIA H100 80GB HBM3", "micro_batch": 16, "accumulation": 1, "effective_global_batch": 64, "supervised_token_loss": true, "NCCL_NVLS_ENABLE": "0", "CUDA_LAUNCH_BLOCKING": "0", "deepspeed": "scripts/zero2_h100.json"}
633
+ [TrainingMemory] {"rank": "3", "stage": "train_begin", "step": 0, "allocated_GiB": 14.601, "reserved_GiB": 19.906, "peak_allocated_GiB": 18.713, "peak_reserved_GiB": 19.906, "device_free_GiB": 57.864, "device_total_GiB": 79.189}
634
+ [TrainingRuntime] {"rank": "2", "versions": {"torch": "2.7.0+cu128", "transformers": "5.3.0", "deepspeed": "0.16.4", "flash-attn": "2.8.3", "triton": "3.3.0", "flash-linear-attention": "0.5.1", "causal-conv1d": "1.6.2.post1"}, "cuda": "12.8", "world_size": 4, "gpu": "NVIDIA H100 80GB HBM3", "micro_batch": 16, "accumulation": 1, "effective_global_batch": 64, "supervised_token_loss": true, "NCCL_NVLS_ENABLE": "0", "CUDA_LAUNCH_BLOCKING": "0", "deepspeed": "scripts/zero2_h100.json"}
635
+ [TrainingMemory] {"rank": "2", "stage": "train_begin", "step": 0, "allocated_GiB": 14.601, "reserved_GiB": 19.906, "peak_allocated_GiB": 18.713, "peak_reserved_GiB": 19.906, "device_free_GiB": 57.864, "device_total_GiB": 79.189}
636
+ [TrainingRuntime] {"rank": "0", "versions": {"torch": "2.7.0+cu128", "transformers": "5.3.0", "deepspeed": "0.16.4", "flash-attn": "2.8.3", "triton": "3.3.0", "flash-linear-attention": "0.5.1", "causal-conv1d": "1.6.2.post1"}, "cuda": "12.8", "world_size": 4, "gpu": "NVIDIA H100 80GB HBM3", "micro_batch": 16, "accumulation": 1, "effective_global_batch": 64, "supervised_token_loss": true, "NCCL_NVLS_ENABLE": "0", "CUDA_LAUNCH_BLOCKING": "0", "deepspeed": "scripts/zero2_h100.json"}
637
+ [TrainingMemory] {"rank": "0", "stage": "train_begin", "step": 0, "allocated_GiB": 14.601, "reserved_GiB": 19.906, "peak_allocated_GiB": 14.601, "peak_reserved_GiB": 19.906, "device_free_GiB": 57.864, "device_total_GiB": 79.189}
638
+
639
  0%| | 0/12 [00:00<?, ?it/s][GeometryBatch] micro_batch=16, mode=per-sample VGGT fallback, input_shapes=[(9, 3, 420, 560), (9, 3, 420, 560), (9, 3, 420, 560), (9, 3, 420, 560), (9, 3, 420, 560), (9, 3, 420, 560), (9, 3, 420, 560), (9, 3, 420, 560), (9, 3, 420, 560), (9, 3, 420, 560), (9, 3, 420, 560), (9, 3, 420, 560), (9, 3, 420, 560), (9, 3, 420, 560), (6, 3, 420, 560), (3, 3, 420, 560)]
640
+ /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.
641
+ with torch.cuda.amp.autocast(dtype=dtype):
642
+ [GeometryBatch] micro_batch=16, mode=per-sample VGGT fallback, input_shapes=[(9, 3, 420, 560), (9, 3, 420, 560), (9, 3, 420, 560), (9, 3, 420, 560), (9, 3, 420, 560), (9, 3, 420, 560), (9, 3, 420, 560), (9, 3, 420, 560), (9, 3, 420, 560), (9, 3, 420, 560), (9, 3, 420, 560), (9, 3, 420, 560), (9, 3, 420, 560), (9, 3, 420, 560), (8, 3, 420, 560), (3, 3, 420, 560)]
643
+ /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.
644
+ with torch.cuda.amp.autocast(dtype=dtype):
645
+ [GeometryBatch] micro_batch=16, mode=per-sample VGGT fallback, input_shapes=[(9, 3, 420, 560), (9, 3, 420, 560), (9, 3, 420, 560), (9, 3, 420, 560), (9, 3, 420, 560), (9, 3, 420, 560), (9, 3, 420, 560), (9, 3, 420, 560), (9, 3, 420, 560), (9, 3, 420, 560), (9, 3, 420, 560), (9, 3, 420, 560), (9, 3, 420, 560), (9, 3, 420, 560), (5, 3, 420, 560), (4, 3, 420, 560)]
646
+ /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.
647
+ with torch.cuda.amp.autocast(dtype=dtype):
648
+ [GeometryBatch] micro_batch=16, mode=per-sample VGGT fallback, input_shapes=[(9, 3, 420, 560), (9, 3, 420, 560), (9, 3, 420, 560), (9, 3, 420, 560), (9, 3, 420, 560), (9, 3, 420, 560), (9, 3, 420, 560), (9, 3, 420, 560), (9, 3, 420, 560), (9, 3, 420, 560), (9, 3, 420, 560), (9, 3, 420, 560), (9, 3, 420, 560), (9, 3, 420, 560), (9, 3, 420, 560), (2, 3, 420, 560)]
649
+ /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.
650
+ with torch.cuda.amp.autocast(dtype=dtype):
651
+ [TrainingMemory] {"rank": "0", "stage": "before_logging", "step": 1, "allocated_GiB": 24.192, "reserved_GiB": 39.621, "peak_allocated_GiB": 45.39, "peak_reserved_GiB": 71.133, "device_free_GiB": 33.303, "device_total_GiB": 79.189}
652
+ [TrainingMemory] {"rank": "2", "stage": "before_logging", "step": 1, "allocated_GiB": 24.172, "reserved_GiB": 71.924, "peak_allocated_GiB": 46.04, "peak_reserved_GiB": 71.924, "device_free_GiB": 1.001, "device_total_GiB": 79.189}
653
+
654
  8%|▊ | 1/12 [01:09<12:49, 69.94s/it][TrainingMemory] {"rank": "1", "stage": "before_logging", "step": 1, "allocated_GiB": 24.172, "reserved_GiB": 71.883, "peak_allocated_GiB": 46.028, "peak_reserved_GiB": 71.883, "device_free_GiB": 1.042, "device_total_GiB": 79.189}
655
+ [TrainingMemory] {"rank": "3", "stage": "before_logging", "step": 1, "allocated_GiB": 24.191, "reserved_GiB": 39.588, "peak_allocated_GiB": 45.972, "peak_reserved_GiB": 68.902, "device_free_GiB": 33.337, "device_total_GiB": 79.189}
656
+ /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.
657
+ with torch.cuda.amp.autocast(dtype=dtype):
658
+ /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.
659
+ with torch.cuda.amp.autocast(dtype=dtype):
660
+ /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.
661
+ with torch.cuda.amp.autocast(dtype=dtype):
662
+ /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.
663
+ with torch.cuda.amp.autocast(dtype=dtype):
664
+ [TrainingMemory] {"rank": "2", "stage": "before_logging", "step": 2, "allocated_GiB": 24.182, "reserved_GiB": 71.924, "peak_allocated_GiB": 54.502, "peak_reserved_GiB": 71.924, "device_free_GiB": 1.001, "device_total_GiB": 79.189}
665
+ [TrainingMemory] {"rank": "3", "stage": "before_logging", "step": 2, "allocated_GiB": 24.173, "reserved_GiB": 63.305, "peak_allocated_GiB": 54.051, "peak_reserved_GiB": 63.305, "device_free_GiB": 9.62, "device_total_GiB": 79.189}
666
+ [TrainingMemory] {"rank": "1", "stage": "before_logging", "step": 2, "allocated_GiB": 24.182, "reserved_GiB": 71.883, "peak_allocated_GiB": 54.57, "peak_reserved_GiB": 71.883, "device_free_GiB": 1.042, "device_total_GiB": 79.189}
667
+ [TrainingMemory] {"rank": "0", "stage": "before_logging", "step": 2, "allocated_GiB": 24.154, "reserved_GiB": 65.676, "peak_allocated_GiB": 53.816, "peak_reserved_GiB": 65.676, "device_free_GiB": 7.247, "device_total_GiB": 79.189}
668
+
669
  17%|█▋ | 2/12 [01:49<08:40, 52.06s/it]/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.
670
+ with torch.cuda.amp.autocast(dtype=dtype):
671
+ [TrainingMemory] {"rank": "1", "stage": "before_logging", "step": 3, "allocated_GiB": 24.229, "reserved_GiB": 71.883, "peak_allocated_GiB": 54.525, "peak_reserved_GiB": 71.883, "device_free_GiB": 1.04, "device_total_GiB": 79.189}
672
+ [TrainingMemory] {"rank": "2", "stage": "before_logging", "step": 3, "allocated_GiB": 24.192, "reserved_GiB": 71.924, "peak_allocated_GiB": 54.035, "peak_reserved_GiB": 71.924, "device_free_GiB": 0.999, "device_total_GiB": 79.189}
673
+ [TrainingMemory] {"rank": "0", "stage": "before_logging", "step": 3, "allocated_GiB": 24.22, "reserved_GiB": 65.676, "peak_allocated_GiB": 54.328, "peak_reserved_GiB": 65.676, "device_free_GiB": 7.247, "device_total_GiB": 79.189}
674
+
675
  25%|██▌ | 3/12 [02:13<05:54, 39.36s/it][TrainingMemory] {"rank": "3", "stage": "before_logging", "step": 3, "allocated_GiB": 24.21, "reserved_GiB": 63.305, "peak_allocated_GiB": 54.124, "peak_reserved_GiB": 63.305, "device_free_GiB": 9.62, "device_total_GiB": 79.189}
676
+ [TrainingMemory] {"rank": "3", "stage": "before_logging", "step": 4, "allocated_GiB": 24.173, "reserved_GiB": 65.004, "peak_allocated_GiB": 54.289, "peak_reserved_GiB": 65.004, "device_free_GiB": 7.919, "device_total_GiB": 79.189}
677
+ [TrainingMemory] {"rank": "1", "stage": "before_logging", "step": 4, "allocated_GiB": 24.163, "reserved_GiB": 71.883, "peak_allocated_GiB": 54.33, "peak_reserved_GiB": 71.883, "device_free_GiB": 1.04, "device_total_GiB": 79.189}
678
+ [TrainingMemory] {"rank": "0", "stage": "before_logging", "step": 4, "allocated_GiB": 24.163, "reserved_GiB": 65.676, "peak_allocated_GiB": 53.39, "peak_reserved_GiB": 65.676, "device_free_GiB": 7.247, "device_total_GiB": 79.189}
679
+ [TrainingMemory] {"rank": "2", "stage": "before_logging", "step": 4, "allocated_GiB": 24.172, "reserved_GiB": 71.924, "peak_allocated_GiB": 54.395, "peak_reserved_GiB": 71.924, "device_free_GiB": 0.999, "device_total_GiB": 79.189}
680
+
681
  33%|███▎ | 4/12 [02:37<04:26, 33.30s/it][TrainingMemory] {"rank": "3", "stage": "before_logging", "step": 5, "allocated_GiB": 24.182, "reserved_GiB": 65.004, "peak_allocated_GiB": 54.512, "peak_reserved_GiB": 65.004, "device_free_GiB": 7.919, "device_total_GiB": 79.189}
682
+ [TrainingMemory] {"rank": "2", "stage": "before_logging", "step": 5, "allocated_GiB": 24.192, "reserved_GiB": 71.924, "peak_allocated_GiB": 54.676, "peak_reserved_GiB": 71.924, "device_free_GiB": 0.999, "device_total_GiB": 79.189}[TrainingMemory] {"rank": "1", "stage": "before_logging", "step": 5, "allocated_GiB": 24.163, "reserved_GiB": 71.883, "peak_allocated_GiB": 54.514, "peak_reserved_GiB": 71.883, "device_free_GiB": 1.04, "device_total_GiB": 79.189}
683
+
684
+ [TrainingMemory] {"rank": "0", "stage": "before_logging", "step": 5, "allocated_GiB": 24.163, "reserved_GiB": 65.676, "peak_allocated_GiB": 53.982, "peak_reserved_GiB": 65.676, "device_free_GiB": 7.247, "device_total_GiB": 79.189}
685
+
686
  42%|████▏ | 5/12 [02:53<03:08, 26.86s/it][TrainingMemory] {"rank": "2", "stage": "before_logging", "step": 6, "allocated_GiB": 24.182, "reserved_GiB": 71.924, "peak_allocated_GiB": 54.164, "peak_reserved_GiB": 71.924, "device_free_GiB": 0.999, "device_total_GiB": 79.189}
687
+ [TrainingMemory] {"rank": "0", "stage": "before_logging", "step": 6, "allocated_GiB": 24.201, "reserved_GiB": 65.676, "peak_allocated_GiB": 53.583, "peak_reserved_GiB": 65.676, "device_free_GiB": 7.247, "device_total_GiB": 79.189}
688
+
689
  50%|█████ | 6/12 [03:08<02:17, 22.90s/it][TrainingMemory] {"rank": "1", "stage": "before_logging", "step": 6, "allocated_GiB": 24.201, "reserved_GiB": 71.883, "peak_allocated_GiB": 54.416, "peak_reserved_GiB": 71.883, "device_free_GiB": 1.04, "device_total_GiB": 79.189}
690
+ [TrainingMemory] {"rank": "3", "stage": "before_logging", "step": 6, "allocated_GiB": 24.192, "reserved_GiB": 65.004, "peak_allocated_GiB": 54.063, "peak_reserved_GiB": 65.004, "device_free_GiB": 7.919, "device_total_GiB": 79.189}
691
+ /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.
692
+ with torch.cuda.amp.autocast(dtype=dtype):
693
+ /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.
694
+ with torch.cuda.amp.autocast(dtype=dtype):
695
+ [TrainingMemory] {"rank": "3", "stage": "before_logging", "step": 7, "allocated_GiB": 24.182, "reserved_GiB": 65.004, "peak_allocated_GiB": 54.513, "peak_reserved_GiB": 65.004, "device_free_GiB": 7.919, "device_total_GiB": 79.189}
696
+ [TrainingMemory] {"rank": "0", "stage": "before_logging", "step": 7, "allocated_GiB": 24.211, "reserved_GiB": 65.676, "peak_allocated_GiB": 53.903, "peak_reserved_GiB": 65.676, "device_free_GiB": 7.247, "device_total_GiB": 79.189}
697
+ [TrainingMemory] {"rank": "2", "stage": "before_logging", "step": 7, "allocated_GiB": 24.22, "reserved_GiB": 71.924, "peak_allocated_GiB": 54.89, "peak_reserved_GiB": 71.924, "device_free_GiB": 0.999, "device_total_GiB": 79.189}
698
+
699
  58%|█████▊ | 7/12 [03:23<01:42, 20.45s/it][TrainingMemory] {"rank": "1", "stage": "before_logging", "step": 7, "allocated_GiB": 24.201, "reserved_GiB": 71.883, "peak_allocated_GiB": 54.855, "peak_reserved_GiB": 71.883, "device_free_GiB": 1.038, "device_total_GiB": 79.189}
700
+ [TrainingMemory] {"rank": "1", "stage": "before_logging", "step": 8, "allocated_GiB": 24.192, "reserved_GiB": 71.883, "peak_allocated_GiB": 55.349, "peak_reserved_GiB": 71.883, "device_free_GiB": 1.038, "device_total_GiB": 79.189}
701
+ [TrainingMemory] {"rank": "0", "stage": "before_logging", "step": 8, "allocated_GiB": 24.191, "reserved_GiB": 66.631, "peak_allocated_GiB": 55.498, "peak_reserved_GiB": 66.631, "device_free_GiB": 6.292, "device_total_GiB": 79.189}
702
+
703
  67%|██████▋ | 8/12 [04:03<01:46, 26.65s/it][TrainingMemory] {"rank": "3", "stage": "before_logging", "step": 8, "allocated_GiB": 24.182, "reserved_GiB": 65.004, "peak_allocated_GiB": 54.43, "peak_reserved_GiB": 65.004, "device_free_GiB": 7.919, "device_total_GiB": 79.189}
704
+ [TrainingMemory] {"rank": "2", "stage": "before_logging", "step": 8, "allocated_GiB": 24.201, "reserved_GiB": 71.924, "peak_allocated_GiB": 55.012, "peak_reserved_GiB": 71.924, "device_free_GiB": 0.999, "device_total_GiB": 79.189}
705
+ /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.
706
+ with torch.cuda.amp.autocast(dtype=dtype):
707
+ [TrainingMemory] {"rank": "1", "stage": "before_logging", "step": 9, "allocated_GiB": 24.22, "reserved_GiB": 71.883, "peak_allocated_GiB": 55.188, "peak_reserved_GiB": 71.883, "device_free_GiB": 1.038, "device_total_GiB": 79.189}[TrainingMemory] {"rank": "2", "stage": "before_logging", "step": 9, "allocated_GiB": 24.229, "reserved_GiB": 71.924, "peak_allocated_GiB": 54.762, "peak_reserved_GiB": 71.924, "device_free_GiB": 0.999, "device_total_GiB": 79.189}
708
+
709
+ [TrainingMemory] {"rank": "0", "stage": "before_logging", "step": 9, "allocated_GiB": 24.182, "reserved_GiB": 66.631, "peak_allocated_GiB": 53.989, "peak_reserved_GiB": 66.631, "device_free_GiB": 6.292, "device_total_GiB": 79.189}
710
+
711
  75%|███████▌ | 9/12 [04:19<01:09, 23.24s/it][TrainingMemory] {"rank": "3", "stage": "before_logging", "step": 9, "allocated_GiB": 24.249, "reserved_GiB": 65.848, "peak_allocated_GiB": 54.738, "peak_reserved_GiB": 65.848, "device_free_GiB": 7.075, "device_total_GiB": 79.189}
712
+ /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.
713
+ with torch.cuda.amp.autocast(dtype=dtype):
714
+ [TrainingMemory] {"rank": "3", "stage": "before_logging", "step": 10, "allocated_GiB": 24.201, "reserved_GiB": 65.848, "peak_allocated_GiB": 54.159, "peak_reserved_GiB": 65.848, "device_free_GiB": 7.075, "device_total_GiB": 79.189}
715
+ [TrainingMemory] {"rank": "2", "stage": "before_logging", "step": 10, "allocated_GiB": 24.192, "reserved_GiB": 71.924, "peak_allocated_GiB": 54.212, "peak_reserved_GiB": 71.924, "device_free_GiB": 0.999, "device_total_GiB": 79.189}
716
+ [TrainingMemory] {"rank": "1", "stage": "before_logging", "step": 10, "allocated_GiB": 24.211, "reserved_GiB": 71.883, "peak_allocated_GiB": 54.644, "peak_reserved_GiB": 71.883, "device_free_GiB": 1.038, "device_total_GiB": 79.189}
717
+ [TrainingMemory] {"rank": "0", "stage": "before_logging", "step": 10, "allocated_GiB": 24.21, "reserved_GiB": 66.631, "peak_allocated_GiB": 53.92, "peak_reserved_GiB": 66.631, "device_free_GiB": 6.29, "device_total_GiB": 79.189}
718
+
719
  83%|████████▎ | 10/12 [04:34<00:41, 20.85s/it][TrainingMemory] {"rank": "1", "stage": "after_logging", "step": 10, "allocated_GiB": 24.211, "reserved_GiB": 71.883, "peak_allocated_GiB": 54.644, "peak_reserved_GiB": 71.883, "device_free_GiB": 0.565, "device_total_GiB": 79.189}
720
+ [TrainingMemory] {"rank": "0", "stage": "after_logging", "step": 10, "allocated_GiB": 24.21, "reserved_GiB": 66.631, "peak_allocated_GiB": 53.92, "peak_reserved_GiB": 66.631, "device_free_GiB": 5.817, "device_total_GiB": 79.189}
721
+ [TrainingMemory] {"rank": "2", "stage": "after_logging", "step": 10, "allocated_GiB": 24.192, "reserved_GiB": 71.924, "peak_allocated_GiB": 54.212, "peak_reserved_GiB": 71.924, "device_free_GiB": 0.526, "device_total_GiB": 79.189}
722
+
723
 
724
+ [TrainingMemory] {"rank": "3", "stage": "after_logging", "step": 10, "allocated_GiB": 24.201, "reserved_GiB": 65.848, "peak_allocated_GiB": 54.159, "peak_reserved_GiB": 65.848, "device_free_GiB": 6.602, "device_total_GiB": 79.189}
725
+
726
  83%|████████▎ | 10/12 [04:35<00:41, 20.85s/it][TrainingMemory] {"rank": "2", "stage": "before_logging", "step": 11, "allocated_GiB": 24.144, "reserved_GiB": 71.924, "peak_allocated_GiB": 54.433, "peak_reserved_GiB": 71.924, "device_free_GiB": 0.524, "device_total_GiB": 79.189}
727
+ [TrainingMemory] {"rank": "1", "stage": "before_logging", "step": 11, "allocated_GiB": 24.144, "reserved_GiB": 71.883, "peak_allocated_GiB": 54.489, "peak_reserved_GiB": 71.883, "device_free_GiB": 0.565, "device_total_GiB": 79.189}
728
+ [TrainingMemory] {"rank": "0", "stage": "before_logging", "step": 11, "allocated_GiB": 24.144, "reserved_GiB": 66.631, "peak_allocated_GiB": 53.685, "peak_reserved_GiB": 66.631, "device_free_GiB": 5.817, "device_total_GiB": 79.189}
729
+
730
  92%|█████████▏| 11/12 [04:50<00:19, 19.24s/it][TrainingMemory] {"rank": "3", "stage": "before_logging", "step": 11, "allocated_GiB": 24.154, "reserved_GiB": 65.848, "peak_allocated_GiB": 54.197, "peak_reserved_GiB": 65.848, "device_free_GiB": 6.602, "device_total_GiB": 79.189}
731
+ [TrainingMemory] {"rank": "3", "stage": "before_logging", "step": 12, "allocated_GiB": 24.191, "reserved_GiB": 65.848, "peak_allocated_GiB": 54.318, "peak_reserved_GiB": 65.848, "device_free_GiB": 6.602, "device_total_GiB": 79.189}
732
+ [TrainingMemory] {"rank": "2", "stage": "before_logging", "step": 12, "allocated_GiB": 24.182, "reserved_GiB": 71.924, "peak_allocated_GiB": 54.444, "peak_reserved_GiB": 71.924, "device_free_GiB": 0.524, "device_total_GiB": 79.189}
733
+ [TrainingMemory] {"rank": "1", "stage": "before_logging", "step": 12, "allocated_GiB": 24.192, "reserved_GiB": 71.883, "peak_allocated_GiB": 54.548, "peak_reserved_GiB": 71.883, "device_free_GiB": 0.565, "device_total_GiB": 79.189}
734
+ [TrainingMemory] {"rank": "0", "stage": "before_logging", "step": 12, "allocated_GiB": 24.173, "reserved_GiB": 66.631, "peak_allocated_GiB": 53.641, "peak_reserved_GiB": 66.631, "device_free_GiB": 5.815, "device_total_GiB": 79.189}
735
+
736
+
737
+
738
+ [TrainingMemory] {"rank": "3", "stage": "after_logging", "step": 12, "allocated_GiB": 22.888, "reserved_GiB": 65.848, "peak_allocated_GiB": 54.318, "peak_reserved_GiB": 65.848, "device_free_GiB": 6.602, "device_total_GiB": 79.189}
739
+ [TrainingMemory] {"rank": "0", "stage": "after_logging", "step": 12, "allocated_GiB": 22.888, "reserved_GiB": 66.631, "peak_allocated_GiB": 53.641, "peak_reserved_GiB": 66.631, "device_free_GiB": 5.815, "device_total_GiB": 79.189}
740
+ [TrainingMemory] {"rank": "1", "stage": "after_logging", "step": 12, "allocated_GiB": 22.888, "reserved_GiB": 71.883, "peak_allocated_GiB": 54.548, "peak_reserved_GiB": 71.883, "device_free_GiB": 0.565, "device_total_GiB": 79.189}
741
+ [TrainingMemory] {"rank": "2", "stage": "after_logging", "step": 12, "allocated_GiB": 22.888, "reserved_GiB": 71.924, "peak_allocated_GiB": 54.444, "peak_reserved_GiB": 71.924, "device_free_GiB": 0.524, "device_total_GiB": 79.189}
742
+
743
 
744
+
745
+
746
+ [rank0]:[W918 03:56:59.052849812 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())