Hugging Face's logo

Datasets:
phamquandung
/
debugging_optimized

Dataset card Files
xet
 Community
phamquandung commited on
Commit
4c78bb9
·
verified ·
1 Parent(s): eeb4795

Upload folder using huggingface_hub

Browse files
Files changed (2) hide show
  1. error_check.91tjCl/bs1/train.log +717 -0
  2. error_check.91tjCl/bs4/train.log +733 -0
error_check.91tjCl/bs1/train.log ADDED
@@ -0,0 +1,717 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
0
  0%| | 0/12 [00:00<?, ?it/s]/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.
 
 
 
 
 
 
 
 
 
 
1
  8%|▊ | 1/12 [01:11<13:02, 71.10s/it][TrainingMemory] {"rank": "2", "stage": "before_logging", "step": 1, "allocated_GiB": 24.172, "reserved_GiB": 41.92, "peak_allocated_GiB": 34.44, "peak_reserved_GiB": 41.92, "device_free_GiB": 31.005, "device_total_GiB": 79.189}
 
 
 
 
 
 
2
  17%|█▋ | 2/12 [01:39<07:37, 45.80s/it][TrainingMemory] {"rank": "3", "stage": "before_logging", "step": 3, "allocated_GiB": 24.182, "reserved_GiB": 41.777, "peak_allocated_GiB": 34.429, "peak_reserved_GiB": 41.777, "device_free_GiB": 31.145, "device_total_GiB": 79.189}
 
 
3
  25%|██▌ | 3/12 [02:07<05:41, 37.96s/it][TrainingMemory] {"rank": "1", "stage": "before_logging", "step": 3, "allocated_GiB": 24.229, "reserved_GiB": 43.025, "peak_allocated_GiB": 34.508, "peak_reserved_GiB": 43.025, "device_free_GiB": 29.897, "device_total_GiB": 79.189}
 
 
 
 
 
4
  33%|███▎ | 4/12 [02:35<04:32, 34.08s/it][TrainingMemory] {"rank": "1", "stage": "before_logging", "step": 4, "allocated_GiB": 24.181, "reserved_GiB": 43.025, "peak_allocated_GiB": 34.461, "peak_reserved_GiB": 43.025, "device_free_GiB": 29.897, "device_total_GiB": 79.189}
 
 
5
  42%|████▏ | 5/12 [03:04<03:43, 31.99s/it][TrainingMemory] {"rank": "2", "stage": "before_logging", "step": 5, "allocated_GiB": 24.163, "reserved_GiB": 41.922, "peak_allocated_GiB": 34.481, "peak_reserved_GiB": 41.922, "device_free_GiB": 31.001, "device_total_GiB": 79.189}
 
 
 
 
 
 
 
6
  50%|█████ | 6/12 [03:32<03:04, 30.78s/it][TrainingMemory] {"rank": "3", "stage": "before_logging", "step": 7, "allocated_GiB": 24.163, "reserved_GiB": 41.777, "peak_allocated_GiB": 34.409, "peak_reserved_GiB": 41.777, "device_free_GiB": 31.145, "device_total_GiB": 79.189}[TrainingMemory] {"rank": "2", "stage": "before_logging", "step": 7, "allocated_GiB": 24.21, "reserved_GiB": 41.922, "peak_allocated_GiB": 34.529, "peak_reserved_GiB": 41.922, "device_free_GiB": 31.001, "device_total_GiB": 79.189}
 
 
 
 
7
  58%|█████▊ | 7/12 [04:00<02:29, 29.96s/it][TrainingMemory] {"rank": "2", "stage": "before_logging", "step": 8, "allocated_GiB": 24.191, "reserved_GiB": 41.922, "peak_allocated_GiB": 34.51, "peak_reserved_GiB": 41.922, "device_free_GiB": 31.001, "device_total_GiB": 79.189}
 
 
 
8
  67%|██████▋ | 8/12 [04:29<01:57, 29.46s/it][TrainingMemory] {"rank": "1", "stage": "before_logging", "step": 8, "allocated_GiB": 24.2, "reserved_GiB": 43.025, "peak_allocated_GiB": 34.479, "peak_reserved_GiB": 43.025, "device_free_GiB": 29.897, "device_total_GiB": 79.189}
 
 
 
9
  75%|███████▌ | 9/12 [04:57<01:27, 29.21s/it][TrainingMemory] {"rank": "2", "stage": "before_logging", "step": 9, "allocated_GiB": 24.219, "reserved_GiB": 41.922, "peak_allocated_GiB": 34.538, "peak_reserved_GiB": 41.922, "device_free_GiB": 31.001, "device_total_GiB": 79.189}
 
 
 
 
 
10
  83%|████████▎ | 10/12 [05:26<00:58, 29.03s/it][TrainingMemory] {"rank": "2", "stage": "before_logging", "step": 10, "allocated_GiB": 24.2, "reserved_GiB": 41.922, "peak_allocated_GiB": 34.519, "peak_reserved_GiB": 41.922, "device_free_GiB": 31.001, "device_total_GiB": 79.189}
 
 
 
 
11
 
 
 
12
  83%|████████▎ | 10/12 [05:27<00:58, 29.03s/it][TrainingMemory] {"rank": "3", "stage": "before_logging", "step": 11, "allocated_GiB": 24.106, "reserved_GiB": 41.777, "peak_allocated_GiB": 34.352, "peak_reserved_GiB": 41.777, "device_free_GiB": 30.671, "device_total_GiB": 79.189}
 
 
 
13
  92%|█████████▏| 11/12 [05:55<00:28, 28.83s/it][TrainingMemory] {"rank": "2", "stage": "before_logging", "step": 11, "allocated_GiB": 24.134, "reserved_GiB": 41.922, "peak_allocated_GiB": 34.453, "peak_reserved_GiB": 41.922, "device_free_GiB": 30.528, "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
+ warmup_ratio is deprecated and will be removed in v5.2. Use `warmup_steps` instead.
10
+ warmup_ratio is deprecated and will be removed in v5.2. Use `warmup_steps` instead.
11
+ Trainer._get_train_sampler replaced with custom implementation.
12
+ [2026-09-18 00:51:25,591] [INFO] [real_accelerator.py:222:get_accelerator] Setting ds_accelerator to cuda (auto detect)
13
+ warmup_ratio is deprecated and will be removed in v5.2. Use `warmup_steps` instead.
14
+ [2026-09-18 00:51:25,600] [INFO] [real_accelerator.py:222:get_accelerator] Setting ds_accelerator to cuda (auto detect)
15
+ [2026-09-18 00:51:25,601] [INFO] [real_accelerator.py:222:get_accelerator] Setting ds_accelerator to cuda (auto detect)
16
+ [2026-09-18 00:51:25,629] [INFO] [real_accelerator.py:222:get_accelerator] Setting ds_accelerator to cuda (auto detect)
17
+ [2026-09-18 00:51:25,898] [INFO] [comm.py:658:init_distributed] cdb=None
18
+ [2026-09-18 00:51:25,901] [INFO] [comm.py:658:init_distributed] cdb=None
19
+ [2026-09-18 00:51:25,901] [INFO] [comm.py:689:init_distributed] Initializing TorchBackend in DeepSpeed with backend nccl
20
+ [2026-09-18 00:51:25,905] [INFO] [comm.py:658:init_distributed] cdb=None
21
+ [2026-09-18 00:51:25,944] [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
+ Qwen3_5Config {
332
+ "architectures": [
333
+ "Qwen3_5ForConditionalGeneration"
334
+ ],
335
+ "dtype": "bfloat16",
336
+ "feature_fusion_method": "deepstack_language_add",
337
+ "fusion_num_layers": 1,
338
+ "geometry_encoder_layers": [
339
+ 11,
340
+ 17,
341
+ 23
342
+ ],
343
+ "geometry_encoder_path": "/mnt/data/vmo-ai-task/dungpq6/model-checkpoint/VGGT-1B",
344
+ "geometry_encoder_streaming": true,
345
+ "geometry_encoder_type": "vggt",
346
+ "geometry_frame_strict": true,
347
+ "geometry_fusion_layers": [
348
+ 0,
349
+ 1,
350
+ 2
351
+ ],
352
+ "geometry_fusion_scale": 0.5,
353
+ "geometry_importance_gate": true,
354
+ "geometry_learnable_scale": true,
355
+ "geometry_merger_type": "mlp",
356
+ "geometry_spatial_bias": false,
357
+ "image_token_id": 248056,
358
+ "include_camera_token": false,
359
+ "model_type": "qwen3_5",
360
+ "pos_encoding_type": "none",
361
+ "reference_frame": "first",
362
+ "supervised_token_loss": true,
363
+ "text_config": {
364
+ "attention_bias": false,
365
+ "attention_dropout": 0.0,
366
+ "attn_output_gate": true,
367
+ "bos_token_id": null,
368
+ "dtype": "bfloat16",
369
+ "eos_token_id": 248044,
370
+ "full_attention_interval": 4,
371
+ "head_dim": 256,
372
+ "hidden_act": "silu",
373
+ "hidden_size": 2560,
374
+ "initializer_range": 0.02,
375
+ "intermediate_size": 9216,
376
+ "layer_types": [
377
+ "linear_attention",
378
+ "linear_attention",
379
+ "linear_attention",
380
+ "full_attention",
381
+ "linear_attention",
382
+ "linear_attention",
383
+ "linear_attention",
384
+ "full_attention",
385
+ "linear_attention",
386
+ "linear_attention",
387
+ "linear_attention",
388
+ "full_attention",
389
+ "linear_attention",
390
+ "linear_attention",
391
+ "linear_attention",
392
+ "full_attention",
393
+ "linear_attention",
394
+ "linear_attention",
395
+ "linear_attention",
396
+ "full_attention",
397
+ "linear_attention",
398
+ "linear_attention",
399
+ "linear_attention",
400
+ "full_attention",
401
+ "linear_attention",
402
+ "linear_attention",
403
+ "linear_attention",
404
+ "full_attention",
405
+ "linear_attention",
406
+ "linear_attention",
407
+ "linear_attention",
408
+ "full_attention"
409
+ ],
410
+ "linear_conv_kernel_dim": 4,
411
+ "linear_key_head_dim": 128,
412
+ "linear_num_key_heads": 16,
413
+ "linear_num_value_heads": 32,
414
+ "linear_value_head_dim": 128,
415
+ "mamba_ssm_dtype": "float32",
416
+ "max_position_embeddings": 262144,
417
+ "mlp_only_layers": [],
418
+ "model_type": "qwen3_5_text",
419
+ "mtp_num_hidden_layers": 1,
420
+ "mtp_use_dedicated_embeddings": false,
421
+ "num_attention_heads": 16,
422
+ "num_hidden_layers": 32,
423
+ "num_key_value_heads": 4,
424
+ "pad_token_id": null,
425
+ "partial_rotary_factor": 0.25,
426
+ "rms_norm_eps": 1e-06,
427
+ "rope_parameters": {
428
+ "mrope_interleaved": true,
429
+ "mrope_section": [
430
+ 11,
431
+ 11,
432
+ 10
433
+ ],
434
+ "partial_rotary_factor": 0.25,
435
+ "rope_theta": 10000000,
436
+ "rope_type": "default"
437
+ },
438
+ "tie_word_embeddings": true,
439
+ "use_cache": true,
440
+ "vocab_size": 248320
441
+ },
442
+ "tie_word_embeddings": true,
443
+ "transformers_version": "5.3.0",
444
+ "use_cache": false,
445
+ "use_geometry_encoder": true,
446
+ "video_token_id": 248057,
447
+ "vision_config": {
448
+ "deepstack_visual_indexes": [],
449
+ "depth": 24,
450
+ "dtype": "bfloat16",
451
+ "hidden_act": "gelu_pytorch_tanh",
452
+ "hidden_size": 1024,
453
+ "in_channels": 3,
454
+ "initializer_range": 0.02,
455
+ "intermediate_size": 4096,
456
+ "model_type": "qwen3_5",
457
+ "num_heads": 16,
458
+ "num_position_embeddings": 2304,
459
+ "out_hidden_size": 2560,
460
+ "patch_size": 16,
461
+ "spatial_merge_size": 2,
462
+ "temporal_patch_size": 2
463
+ },
464
+ "vision_end_token_id": 248054,
465
+ "vision_language_fusion_layers": null,
466
+ "vision_start_token_id": 248053
467
+ }
468
+
469
+ 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'}]
470
+ >>>>> STOP loss up-weight: 3.0 on token ids [48964]
471
+ Vision Module - Attention Blocks:
472
+ Trainable Block Indices: None
473
+ 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]
474
+ Merger Module Trainable: True
475
+ LLM Module - Embed Tokens Trainable: True
476
+ 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]
477
+ LLM Module - Non-Trainable Layer Indices: None
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
+
631
+ [2026-09-18 00:52:23] worker-1:2222578:2223325 [0] ras/client_support.cc:160 NCCL WARN Call to bind failed: Address already in use
632
+
633
+ [2026-09-18 00:52:23] worker-1:2222580:2223331 [2] ras/client_support.cc:160 NCCL WARN Call to bind failed: Address already in use
634
+
635
+ [2026-09-18 00:52:23] worker-1:2222581:2223326 [3] ras/client_support.cc:160 NCCL WARN Call to bind failed: Address already in use
636
+
637
+ [2026-09-18 00:52:23] worker-1:2222579:2223327 [1] ras/client_support.cc:160 NCCL WARN Call to bind failed: Address already in use
638
+ [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": 1, "accumulation": 16, "effective_global_batch": 64, "supervised_token_loss": true, "NCCL_NVLS_ENABLE": "0", "CUDA_LAUNCH_BLOCKING": "1", "deepspeed": "scripts/zero2_h100.json"}
639
+ [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}
640
+ [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": 1, "accumulation": 16, "effective_global_batch": 64, "supervised_token_loss": true, "NCCL_NVLS_ENABLE": "0", "CUDA_LAUNCH_BLOCKING": "1", "deepspeed": "scripts/zero2_h100.json"}
641
+ [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}
642
+ [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": 1, "accumulation": 16, "effective_global_batch": 64, "supervised_token_loss": true, "NCCL_NVLS_ENABLE": "0", "CUDA_LAUNCH_BLOCKING": "1", "deepspeed": "scripts/zero2_h100.json"}
643
+ [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}
644
+ [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": 1, "accumulation": 16, "effective_global_batch": 64, "supervised_token_loss": true, "NCCL_NVLS_ENABLE": "0", "CUDA_LAUNCH_BLOCKING": "1", "deepspeed": "scripts/zero2_h100.json"}
645
+ [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}
646
+
647
  0%| | 0/12 [00:00<?, ?it/s]/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.
648
+ with torch.cuda.amp.autocast(dtype=dtype):
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
+ /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.
652
+ with torch.cuda.amp.autocast(dtype=dtype):
653
+ /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.
654
+ with torch.cuda.amp.autocast(dtype=dtype):
655
+ [TrainingMemory] {"rank": "0", "stage": "before_logging", "step": 1, "allocated_GiB": 24.21, "reserved_GiB": 43.633, "peak_allocated_GiB": 34.419, "peak_reserved_GiB": 43.633, "device_free_GiB": 29.292, "device_total_GiB": 79.189}
656
+ [TrainingMemory] {"rank": "1", "stage": "before_logging", "step": 1, "allocated_GiB": 24.191, "reserved_GiB": 43.023, "peak_allocated_GiB": 34.419, "peak_reserved_GiB": 43.023, "device_free_GiB": 29.901, "device_total_GiB": 79.189}
657
+
658
  8%|▊ | 1/12 [01:11<13:02, 71.10s/it][TrainingMemory] {"rank": "2", "stage": "before_logging", "step": 1, "allocated_GiB": 24.172, "reserved_GiB": 41.92, "peak_allocated_GiB": 34.44, "peak_reserved_GiB": 41.92, "device_free_GiB": 31.005, "device_total_GiB": 79.189}
659
+ [TrainingMemory] {"rank": "3", "stage": "before_logging", "step": 1, "allocated_GiB": 24.153, "reserved_GiB": 41.775, "peak_allocated_GiB": 34.349, "peak_reserved_GiB": 41.775, "device_free_GiB": 31.149, "device_total_GiB": 79.189}
660
+ [TrainingMemory] {"rank": "2", "stage": "before_logging", "step": 2, "allocated_GiB": 24.163, "reserved_GiB": 41.922, "peak_allocated_GiB": 34.482, "peak_reserved_GiB": 41.922, "device_free_GiB": 31.001, "device_total_GiB": 79.189}
661
+ [TrainingMemory] {"rank": "0", "stage": "before_logging", "step": 2, "allocated_GiB": 24.2, "reserved_GiB": 43.633, "peak_allocated_GiB": 34.46, "peak_reserved_GiB": 43.633, "device_free_GiB": 29.292, "device_total_GiB": 79.189}
662
+ [TrainingMemory] {"rank": "3", "stage": "before_logging", "step": 2, "allocated_GiB": 24.134, "reserved_GiB": 41.777, "peak_allocated_GiB": 34.381, "peak_reserved_GiB": 41.777, "device_free_GiB": 31.145, "device_total_GiB": 79.189}
663
+ [TrainingMemory] {"rank": "1", "stage": "before_logging", "step": 2, "allocated_GiB": 24.191, "reserved_GiB": 43.023, "peak_allocated_GiB": 34.471, "peak_reserved_GiB": 43.023, "device_free_GiB": 29.901, "device_total_GiB": 79.189}
664
+
665
  17%|█▋ | 2/12 [01:39<07:37, 45.80s/it][TrainingMemory] {"rank": "3", "stage": "before_logging", "step": 3, "allocated_GiB": 24.182, "reserved_GiB": 41.777, "peak_allocated_GiB": 34.429, "peak_reserved_GiB": 41.777, "device_free_GiB": 31.145, "device_total_GiB": 79.189}
666
+ [TrainingMemory] {"rank": "0", "stage": "before_logging", "step": 3, "allocated_GiB": 24.238, "reserved_GiB": 43.635, "peak_allocated_GiB": 34.498, "peak_reserved_GiB": 43.635, "device_free_GiB": 29.29, "device_total_GiB": 79.189}
667
+
668
  25%|██▌ | 3/12 [02:07<05:41, 37.96s/it][TrainingMemory] {"rank": "1", "stage": "before_logging", "step": 3, "allocated_GiB": 24.229, "reserved_GiB": 43.025, "peak_allocated_GiB": 34.508, "peak_reserved_GiB": 43.025, "device_free_GiB": 29.897, "device_total_GiB": 79.189}
669
+ [TrainingMemory] {"rank": "2", "stage": "before_logging", "step": 3, "allocated_GiB": 24.2, "reserved_GiB": 41.922, "peak_allocated_GiB": 34.519, "peak_reserved_GiB": 41.922, "device_free_GiB": 31.001, "device_total_GiB": 79.189}
670
+ [TrainingMemory] {"rank": "3", "stage": "before_logging", "step": 4, "allocated_GiB": 24.125, "reserved_GiB": 41.777, "peak_allocated_GiB": 34.371, "peak_reserved_GiB": 41.777, "device_free_GiB": 31.145, "device_total_GiB": 79.189}
671
+ [TrainingMemory] {"rank": "2", "stage": "before_logging", "step": 4, "allocated_GiB": 24.172, "reserved_GiB": 41.922, "peak_allocated_GiB": 34.491, "peak_reserved_GiB": 41.922, "device_free_GiB": 31.001, "device_total_GiB": 79.189}
672
+ [TrainingMemory] {"rank": "0", "stage": "before_logging", "step": 4, "allocated_GiB": 24.191, "reserved_GiB": 43.635, "peak_allocated_GiB": 34.45, "peak_reserved_GiB": 43.635, "device_free_GiB": 29.29, "device_total_GiB": 79.189}
673
+
674
  33%|███▎ | 4/12 [02:35<04:32, 34.08s/it][TrainingMemory] {"rank": "1", "stage": "before_logging", "step": 4, "allocated_GiB": 24.181, "reserved_GiB": 43.025, "peak_allocated_GiB": 34.461, "peak_reserved_GiB": 43.025, "device_free_GiB": 29.897, "device_total_GiB": 79.189}
675
+ [TrainingMemory] {"rank": "0", "stage": "before_logging", "step": 5, "allocated_GiB": 24.21, "reserved_GiB": 43.635, "peak_allocated_GiB": 34.469, "peak_reserved_GiB": 43.635, "device_free_GiB": 29.29, "device_total_GiB": 79.189}
676
+
677
  42%|████▏ | 5/12 [03:04<03:43, 31.99s/it][TrainingMemory] {"rank": "2", "stage": "before_logging", "step": 5, "allocated_GiB": 24.163, "reserved_GiB": 41.922, "peak_allocated_GiB": 34.481, "peak_reserved_GiB": 41.922, "device_free_GiB": 31.001, "device_total_GiB": 79.189}
678
+ [TrainingMemory] {"rank": "1", "stage": "before_logging", "step": 5, "allocated_GiB": 24.181, "reserved_GiB": 43.025, "peak_allocated_GiB": 34.46, "peak_reserved_GiB": 43.025, "device_free_GiB": 29.897, "device_total_GiB": 79.189}
679
+ [TrainingMemory] {"rank": "3", "stage": "before_logging", "step": 5, "allocated_GiB": 24.144, "reserved_GiB": 41.777, "peak_allocated_GiB": 34.39, "peak_reserved_GiB": 41.777, "device_free_GiB": 31.145, "device_total_GiB": 79.189}
680
+ [TrainingMemory] {"rank": "1", "stage": "before_logging", "step": 6, "allocated_GiB": 24.2, "reserved_GiB": 43.025, "peak_allocated_GiB": 34.479, "peak_reserved_GiB": 43.025, "device_free_GiB": 29.897, "device_total_GiB": 79.189}
681
+ [TrainingMemory] {"rank": "2", "stage": "before_logging", "step": 6, "allocated_GiB": 24.191, "reserved_GiB": 41.922, "peak_allocated_GiB": 34.51, "peak_reserved_GiB": 41.922, "device_free_GiB": 31.001, "device_total_GiB": 79.189}
682
+ [TrainingMemory] {"rank": "3", "stage": "before_logging", "step": 6, "allocated_GiB": 24.162, "reserved_GiB": 41.777, "peak_allocated_GiB": 34.409, "peak_reserved_GiB": 41.777, "device_free_GiB": 31.145, "device_total_GiB": 79.189}
683
+ [TrainingMemory] {"rank": "0", "stage": "before_logging", "step": 6, "allocated_GiB": 24.219, "reserved_GiB": 43.635, "peak_allocated_GiB": 34.479, "peak_reserved_GiB": 43.635, "device_free_GiB": 29.288, "device_total_GiB": 79.189}
684
+
685
  50%|█████ | 6/12 [03:32<03:04, 30.78s/it][TrainingMemory] {"rank": "3", "stage": "before_logging", "step": 7, "allocated_GiB": 24.163, "reserved_GiB": 41.777, "peak_allocated_GiB": 34.409, "peak_reserved_GiB": 41.777, "device_free_GiB": 31.145, "device_total_GiB": 79.189}[TrainingMemory] {"rank": "2", "stage": "before_logging", "step": 7, "allocated_GiB": 24.21, "reserved_GiB": 41.922, "peak_allocated_GiB": 34.529, "peak_reserved_GiB": 41.922, "device_free_GiB": 31.001, "device_total_GiB": 79.189}
686
+
687
+ [TrainingMemory] {"rank": "1", "stage": "before_logging", "step": 7, "allocated_GiB": 24.219, "reserved_GiB": 43.025, "peak_allocated_GiB": 34.498, "peak_reserved_GiB": 43.025, "device_free_GiB": 29.897, "device_total_GiB": 79.189}
688
+ [TrainingMemory] {"rank": "0", "stage": "before_logging", "step": 7, "allocated_GiB": 24.219, "reserved_GiB": 43.635, "peak_allocated_GiB": 34.479, "peak_reserved_GiB": 43.635, "device_free_GiB": 29.288, "device_total_GiB": 79.189}
689
+
690
  58%|█████▊ | 7/12 [04:00<02:29, 29.96s/it][TrainingMemory] {"rank": "2", "stage": "before_logging", "step": 8, "allocated_GiB": 24.191, "reserved_GiB": 41.922, "peak_allocated_GiB": 34.51, "peak_reserved_GiB": 41.922, "device_free_GiB": 31.001, "device_total_GiB": 79.189}
691
+ [TrainingMemory] {"rank": "3", "stage": "before_logging", "step": 8, "allocated_GiB": 24.153, "reserved_GiB": 41.777, "peak_allocated_GiB": 34.4, "peak_reserved_GiB": 41.777, "device_free_GiB": 31.145, "device_total_GiB": 79.189}
692
+ [TrainingMemory] {"rank": "0", "stage": "before_logging", "step": 8, "allocated_GiB": 24.219, "reserved_GiB": 43.635, "peak_allocated_GiB": 34.479, "peak_reserved_GiB": 43.635, "device_free_GiB": 29.288, "device_total_GiB": 79.189}
693
+
694
  67%|██████▋ | 8/12 [04:29<01:57, 29.46s/it][TrainingMemory] {"rank": "1", "stage": "before_logging", "step": 8, "allocated_GiB": 24.2, "reserved_GiB": 43.025, "peak_allocated_GiB": 34.479, "peak_reserved_GiB": 43.025, "device_free_GiB": 29.897, "device_total_GiB": 79.189}
695
+ [TrainingMemory] {"rank": "0", "stage": "before_logging", "step": 9, "allocated_GiB": 24.248, "reserved_GiB": 43.635, "peak_allocated_GiB": 34.508, "peak_reserved_GiB": 43.635, "device_free_GiB": 29.288, "device_total_GiB": 79.189}
696
+ [TrainingMemory] {"rank": "3", "stage": "before_logging", "step": 9, "allocated_GiB": 24.182, "reserved_GiB": 41.777, "peak_allocated_GiB": 34.428, "peak_reserved_GiB": 41.777, "device_free_GiB": 31.145, "device_total_GiB": 79.189}
697
+
698
  75%|███████▌ | 9/12 [04:57<01:27, 29.21s/it][TrainingMemory] {"rank": "2", "stage": "before_logging", "step": 9, "allocated_GiB": 24.219, "reserved_GiB": 41.922, "peak_allocated_GiB": 34.538, "peak_reserved_GiB": 41.922, "device_free_GiB": 31.001, "device_total_GiB": 79.189}
699
+ [TrainingMemory] {"rank": "1", "stage": "before_logging", "step": 9, "allocated_GiB": 24.229, "reserved_GiB": 43.025, "peak_allocated_GiB": 34.508, "peak_reserved_GiB": 43.025, "device_free_GiB": 29.897, "device_total_GiB": 79.189}
700
+ [TrainingMemory] {"rank": "3", "stage": "before_logging", "step": 10, "allocated_GiB": 24.172, "reserved_GiB": 41.777, "peak_allocated_GiB": 34.419, "peak_reserved_GiB": 41.777, "device_free_GiB": 31.145, "device_total_GiB": 79.189}
701
+ [TrainingMemory] {"rank": "1", "stage": "before_logging", "step": 10, "allocated_GiB": 24.21, "reserved_GiB": 43.025, "peak_allocated_GiB": 34.489, "peak_reserved_GiB": 43.025, "device_free_GiB": 29.897, "device_total_GiB": 79.189}
702
+ [TrainingMemory] {"rank": "0", "stage": "before_logging", "step": 10, "allocated_GiB": 24.229, "reserved_GiB": 43.635, "peak_allocated_GiB": 34.488, "peak_reserved_GiB": 43.635, "device_free_GiB": 29.288, "device_total_GiB": 79.189}
703
+
704
  83%|████████▎ | 10/12 [05:26<00:58, 29.03s/it][TrainingMemory] {"rank": "2", "stage": "before_logging", "step": 10, "allocated_GiB": 24.2, "reserved_GiB": 41.922, "peak_allocated_GiB": 34.519, "peak_reserved_GiB": 41.922, "device_free_GiB": 31.001, "device_total_GiB": 79.189}
705
+ [TrainingMemory] {"rank": "0", "stage": "after_logging", "step": 10, "allocated_GiB": 24.229, "reserved_GiB": 43.635, "peak_allocated_GiB": 34.488, "peak_reserved_GiB": 43.635, "device_free_GiB": 28.815, "device_total_GiB": 79.189}[TrainingMemory] {"rank": "1", "stage": "after_logging", "step": 10, "allocated_GiB": 24.21, "reserved_GiB": 43.025, "peak_allocated_GiB": 34.489, "peak_reserved_GiB": 43.025, "device_free_GiB": 29.425, "device_total_GiB": 79.189}
706
+
707
+ [TrainingMemory] {"rank": "3", "stage": "after_logging", "step": 10, "allocated_GiB": 24.172, "reserved_GiB": 41.777, "peak_allocated_GiB": 34.419, "peak_reserved_GiB": 41.777, "device_free_GiB": 30.673, "device_total_GiB": 79.189}
708
+
709
 
710
+ [TrainingMemory] {"rank": "2", "stage": "after_logging", "step": 10, "allocated_GiB": 24.2, "reserved_GiB": 41.922, "peak_allocated_GiB": 34.519, "peak_reserved_GiB": 41.922, "device_free_GiB": 30.528, "device_total_GiB": 79.189}
711
+
712
  83%|████████▎ | 10/12 [05:27<00:58, 29.03s/it][TrainingMemory] {"rank": "3", "stage": "before_logging", "step": 11, "allocated_GiB": 24.106, "reserved_GiB": 41.777, "peak_allocated_GiB": 34.352, "peak_reserved_GiB": 41.777, "device_free_GiB": 30.671, "device_total_GiB": 79.189}
713
+ [TrainingMemory] {"rank": "0", "stage": "before_logging", "step": 11, "allocated_GiB": 24.181, "reserved_GiB": 43.635, "peak_allocated_GiB": 34.441, "peak_reserved_GiB": 43.635, "device_free_GiB": 28.815, "device_total_GiB": 79.189}
714
+ [TrainingMemory] {"rank": "1", "stage": "before_logging", "step": 11, "allocated_GiB": 24.162, "reserved_GiB": 43.025, "peak_allocated_GiB": 34.441, "peak_reserved_GiB": 43.025, "device_free_GiB": 29.425, "device_total_GiB": 79.189}
715
+
716
  92%|█████████▏| 11/12 [05:55<00:28, 28.83s/it][TrainingMemory] {"rank": "2", "stage": "before_logging", "step": 11, "allocated_GiB": 24.134, "reserved_GiB": 41.922, "peak_allocated_GiB": 34.453, "peak_reserved_GiB": 41.922, "device_free_GiB": 30.528, "device_total_GiB": 79.189}
717
+ [TrainingMemory] {"rank": "2", "stage": "before_logging", "step": 12, "allocated_GiB": 24.181, "reserved_GiB": 41.922, "peak_allocated_GiB": 34.5, "peak_reserved_GiB": 41.922, "device_free_GiB": 30.528, "device_total_GiB": 79.189}
718
+ [TrainingMemory] {"rank": "3", "stage": "before_logging", "step": 12, "allocated_GiB": 24.153, "reserved_GiB": 41.777, "peak_allocated_GiB": 34.4, "peak_reserved_GiB": 41.777, "device_free_GiB": 30.671, "device_total_GiB": 79.189}
719
+ [TrainingMemory] {"rank": "1", "stage": "before_logging", "step": 12, "allocated_GiB": 24.2, "reserved_GiB": 43.025, "peak_allocated_GiB": 34.479, "peak_reserved_GiB": 43.025, "device_free_GiB": 29.425, "device_total_GiB": 79.189}
720
+ [TrainingMemory] {"rank": "0", "stage": "before_logging", "step": 12, "allocated_GiB": 24.2, "reserved_GiB": 43.635, "peak_allocated_GiB": 34.46, "peak_reserved_GiB": 43.635, "device_free_GiB": 28.815, "device_total_GiB": 79.189}
721
+
722
+
723
+
724
+ [TrainingMemory] {"rank": "0", "stage": "after_logging", "step": 12, "allocated_GiB": 22.888, "reserved_GiB": 43.635, "peak_allocated_GiB": 34.46, "peak_reserved_GiB": 43.635, "device_free_GiB": 28.815, "device_total_GiB": 79.189}
725
+ [TrainingMemory] {"rank": "1", "stage": "after_logging", "step": 12, "allocated_GiB": 22.888, "reserved_GiB": 43.025, "peak_allocated_GiB": 34.479, "peak_reserved_GiB": 43.025, "device_free_GiB": 29.425, "device_total_GiB": 79.189}
726
+
727
 
728
+ [TrainingMemory] {"rank": "2", "stage": "after_logging", "step": 12, "allocated_GiB": 22.888, "reserved_GiB": 41.922, "peak_allocated_GiB": 34.5, "peak_reserved_GiB": 41.922, "device_free_GiB": 30.528, "device_total_GiB": 79.189}
729
+
730
+
731
+
732
+ [rank0]:[W918 01:00:24.702458828 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())
error_check.91tjCl/bs4/train.log ADDED
@@ -0,0 +1,733 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
0
  0%| | 0/12 [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)]
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
  8%|▊ | 1/12 [00:50<09:13, 50.31s/it][TrainingMemory] {"rank": "1", "stage": "before_logging", "step": 1, "allocated_GiB": 24.172, "reserved_GiB": 47.205, "peak_allocated_GiB": 34.4, "peak_reserved_GiB": 47.205, "device_free_GiB": 25.747, "device_total_GiB": 79.189}
 
 
 
 
 
2
  17%|█▋ | 2/12 [01:04<04:51, 29.11s/it][TrainingMemory] {"rank": "3", "stage": "before_logging", "step": 3, "allocated_GiB": 24.229, "reserved_GiB": 47.088, "peak_allocated_GiB": 36.386, "peak_reserved_GiB": 47.088, "device_free_GiB": 25.866, "device_total_GiB": 79.189}[TrainingMemory] {"rank": "2", "stage": "before_logging", "step": 3, "allocated_GiB": 24.257, "reserved_GiB": 47.305, "peak_allocated_GiB": 36.563, "peak_reserved_GiB": 47.305, "device_free_GiB": 25.647, "device_total_GiB": 79.189}
 
 
 
 
3
  25%|██▌ | 3/12 [01:19<03:24, 22.69s/it][TrainingMemory] {"rank": "1", "stage": "before_logging", "step": 4, "allocated_GiB": 24.173, "reserved_GiB": 47.205, "peak_allocated_GiB": 36.513, "peak_reserved_GiB": 47.205, "device_free_GiB": 25.747, "device_total_GiB": 79.189}
 
 
 
 
4
  33%|███▎ | 4/12 [01:43<03:05, 23.13s/it][TrainingMemory] {"rank": "1", "stage": "before_logging", "step": 5, "allocated_GiB": 24.163, "reserved_GiB": 47.205, "peak_allocated_GiB": 36.413, "peak_reserved_GiB": 47.205, "device_free_GiB": 25.747, "device_total_GiB": 79.189}
 
 
 
 
5
  42%|████▏ | 5/12 [01:57<02:19, 19.98s/it][TrainingMemory] {"rank": "3", "stage": "before_logging", "step": 6, "allocated_GiB": 24.238, "reserved_GiB": 47.088, "peak_allocated_GiB": 36.414, "peak_reserved_GiB": 47.088, "device_free_GiB": 25.864, "device_total_GiB": 79.189}
 
 
 
 
6
  50%|█████ | 6/12 [02:12<01:49, 18.26s/it][TrainingMemory] {"rank": "3", "stage": "before_logging", "step": 7, "allocated_GiB": 24.219, "reserved_GiB": 47.088, "peak_allocated_GiB": 36.453, "peak_reserved_GiB": 47.088, "device_free_GiB": 25.864, "device_total_GiB": 79.189}
 
 
7
  58%|█████▊ | 7/12 [02:27<01:25, 17.03s/it][TrainingMemory] {"rank": "2", "stage": "before_logging", "step": 7, "allocated_GiB": 24.238, "reserved_GiB": 47.305, "peak_allocated_GiB": 36.63, "peak_reserved_GiB": 47.305, "device_free_GiB": 25.647, "device_total_GiB": 79.189}
 
 
 
 
 
8
  67%|██████▋ | 8/12 [02:42<01:06, 16.57s/it][TrainingMemory] {"rank": "2", "stage": "before_logging", "step": 8, "allocated_GiB": 24.191, "reserved_GiB": 47.305, "peak_allocated_GiB": 36.538, "peak_reserved_GiB": 47.305, "device_free_GiB": 25.647, "device_total_GiB": 79.189}
 
 
 
 
9
  75%|███████▌ | 9/12 [02:56<00:47, 15.75s/it][TrainingMemory] {"rank": "3", "stage": "before_logging", "step": 9, "allocated_GiB": 24.248, "reserved_GiB": 47.088, "peak_allocated_GiB": 36.51, "peak_reserved_GiB": 47.088, "device_free_GiB": 25.864, "device_total_GiB": 79.189}
 
 
10
  83%|████████▎ | 10/12 [03:12<00:31, 15.69s/it][TrainingMemory] {"rank": "3", "stage": "before_logging", "step": 10, "allocated_GiB": 24.21, "reserved_GiB": 47.088, "peak_allocated_GiB": 36.379, "peak_reserved_GiB": 47.088, "device_free_GiB": 25.864, "device_total_GiB": 79.189}
 
 
 
 
 
 
 
11
 
 
12
  83%|████████▎ | 10/12 [03:12<00:31, 15.69s/it][TrainingMemory] {"rank": "1", "stage": "before_logging", "step": 11, "allocated_GiB": 24.162, "reserved_GiB": 47.205, "peak_allocated_GiB": 36.394, "peak_reserved_GiB": 47.205, "device_free_GiB": 25.274, "device_total_GiB": 79.189}
 
 
13
  92%|█████████▏| 11/12 [03:27<00:15, 15.68s/it][TrainingMemory] {"rank": "3", "stage": "before_logging", "step": 11, "allocated_GiB": 24.153, "reserved_GiB": 47.088, "peak_allocated_GiB": 36.366, "peak_reserved_GiB": 47.088, "device_free_GiB": 25.391, "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 01:00:34,676] [INFO] [real_accelerator.py:222:get_accelerator] Setting ds_accelerator to cuda (auto detect)
14
+ [2026-09-18 01:00:34,694] [INFO] [real_accelerator.py:222:get_accelerator] Setting ds_accelerator to cuda (auto detect)
15
+ [2026-09-18 01:00:34,694] [INFO] [real_accelerator.py:222:get_accelerator] Setting ds_accelerator to cuda (auto detect)
16
+ [2026-09-18 01:00:34,695] [INFO] [real_accelerator.py:222:get_accelerator] Setting ds_accelerator to cuda (auto detect)
17
+ [2026-09-18 01:00:34,987] [INFO] [comm.py:658:init_distributed] cdb=None
18
+ [2026-09-18 01:00:34,987] [INFO] [comm.py:658:init_distributed] cdb=None
19
+ [2026-09-18 01:00:34,997] [INFO] [comm.py:658:init_distributed] cdb=None
20
+ [2026-09-18 01:00:34,997] [INFO] [comm.py:658:init_distributed] cdb=None
21
+ [2026-09-18 01:00:34,998] [INFO] [comm.py:689:init_distributed] Initializing TorchBackend in DeepSpeed with backend nccl
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
+ Vision Module - Attention Blocks:
192
+ Trainable Block Indices: None
193
+ 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]
194
+ Merger Module Trainable: True
195
+ LLM Module - Embed Tokens Trainable: True
196
+ 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]
197
+ LLM Module - Non-Trainable Layer Indices: None
198
+ Qwen3_5Config {
199
+ "architectures": [
200
+ "Qwen3_5ForConditionalGeneration"
201
+ ],
202
+ "dtype": "bfloat16",
203
+ "feature_fusion_method": "deepstack_language_add",
204
+ "fusion_num_layers": 1,
205
+ "geometry_encoder_layers": [
206
+ 11,
207
+ 17,
208
+ 23
209
+ ],
210
+ "geometry_encoder_path": "/mnt/data/vmo-ai-task/dungpq6/model-checkpoint/VGGT-1B",
211
+ "geometry_encoder_streaming": true,
212
+ "geometry_encoder_type": "vggt",
213
+ "geometry_frame_strict": true,
214
+ "geometry_fusion_layers": [
215
+ 0,
216
+ 1,
217
+ 2
218
+ ],
219
+ "geometry_fusion_scale": 0.5,
220
+ "geometry_importance_gate": true,
221
+ "geometry_learnable_scale": true,
222
+ "geometry_merger_type": "mlp",
223
+ "geometry_spatial_bias": false,
224
+ "image_token_id": 248056,
225
+ "include_camera_token": false,
226
+ "model_type": "qwen3_5",
227
+ "pos_encoding_type": "none",
228
+ "reference_frame": "first",
229
+ "supervised_token_loss": true,
230
+ "text_config": {
231
+ "attention_bias": false,
232
+ "attention_dropout": 0.0,
233
+ "attn_output_gate": true,
234
+ "bos_token_id": null,
235
+ "dtype": "bfloat16",
236
+ "eos_token_id": 248044,
237
+ "full_attention_interval": 4,
238
+ "head_dim": 256,
239
+ "hidden_act": "silu",
240
+ "hidden_size": 2560,
241
+ "initializer_range": 0.02,
242
+ "intermediate_size": 9216,
243
+ "layer_types": [
244
+ "linear_attention",
245
+ "linear_attention",
246
+ "linear_attention",
247
+ "full_attention",
248
+ "linear_attention",
249
+ "linear_attention",
250
+ "linear_attention",
251
+ "full_attention",
252
+ "linear_attention",
253
+ "linear_attention",
254
+ "linear_attention",
255
+ "full_attention",
256
+ "linear_attention",
257
+ "linear_attention",
258
+ "linear_attention",
259
+ "full_attention",
260
+ "linear_attention",
261
+ "linear_attention",
262
+ "linear_attention",
263
+ "full_attention",
264
+ "linear_attention",
265
+ "linear_attention",
266
+ "linear_attention",
267
+ "full_attention",
268
+ "linear_attention",
269
+ "linear_attention",
270
+ "linear_attention",
271
+ "full_attention",
272
+ "linear_attention",
273
+ "linear_attention",
274
+ "linear_attention",
275
+ "full_attention"
276
+ ],
277
+ "linear_conv_kernel_dim": 4,
278
+ "linear_key_head_dim": 128,
279
+ "linear_num_key_heads": 16,
280
+ "linear_num_value_heads": 32,
281
+ "linear_value_head_dim": 128,
282
+ "mamba_ssm_dtype": "float32",
283
+ "max_position_embeddings": 262144,
284
+ "mlp_only_layers": [],
285
+ "model_type": "qwen3_5_text",
286
+ "mtp_num_hidden_layers": 1,
287
+ "mtp_use_dedicated_embeddings": false,
288
+ "num_attention_heads": 16,
289
+ "num_hidden_layers": 32,
290
+ "num_key_value_heads": 4,
291
+ "pad_token_id": null,
292
+ "partial_rotary_factor": 0.25,
293
+ "rms_norm_eps": 1e-06,
294
+ "rope_parameters": {
295
+ "mrope_interleaved": true,
296
+ "mrope_section": [
297
+ 11,
298
+ 11,
299
+ 10
300
+ ],
301
+ "partial_rotary_factor": 0.25,
302
+ "rope_theta": 10000000,
303
+ "rope_type": "default"
304
+ },
305
+ "tie_word_embeddings": true,
306
+ "use_cache": true,
307
+ "vocab_size": 248320
308
+ },
309
+ "tie_word_embeddings": true,
310
+ "transformers_version": "5.3.0",
311
+ "use_cache": false,
312
+ "use_geometry_encoder": true,
313
+ "video_token_id": 248057,
314
+ "vision_config": {
315
+ "deepstack_visual_indexes": [],
316
+ "depth": 24,
317
+ "dtype": "bfloat16",
318
+ "hidden_act": "gelu_pytorch_tanh",
319
+ "hidden_size": 1024,
320
+ "in_channels": 3,
321
+ "initializer_range": 0.02,
322
+ "intermediate_size": 4096,
323
+ "model_type": "qwen3_5",
324
+ "num_heads": 16,
325
+ "num_position_embeddings": 2304,
326
+ "out_hidden_size": 2560,
327
+ "patch_size": 16,
328
+ "spatial_merge_size": 2,
329
+ "temporal_patch_size": 2
330
+ },
331
+ "vision_end_token_id": 248054,
332
+ "vision_language_fusion_layers": null,
333
+ "vision_start_token_id": 248053
334
+ }
335
+
336
+ 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'}]
337
+ >>>>> STOP loss up-weight: 3.0 on token ids [48964]
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
+
631
+ [2026-09-18 01:01:33] worker-1:2223885:2224630 [0] ras/client_support.cc:160 NCCL WARN Call to bind failed: Address already in use
632
+
633
+ [2026-09-18 01:01:33] worker-1:2223886:2224632 [1] ras/client_support.cc:160 NCCL WARN Call to bind failed: Address already in use
634
+
635
+ [2026-09-18 01:01:33] worker-1:2223888:2224631 [3] ras/client_support.cc:160 NCCL WARN Call to bind failed: Address already in use
636
+
637
+ [2026-09-18 01:01:33] worker-1:2223887:2224633 [2] ras/client_support.cc:160 NCCL WARN Call to bind failed: Address already in use
638
+ [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": 4, "accumulation": 4, "effective_global_batch": 64, "supervised_token_loss": true, "NCCL_NVLS_ENABLE": "0", "CUDA_LAUNCH_BLOCKING": "0", "deepspeed": "scripts/zero2_h100.json"}
639
+ [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}
640
+ [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": 4, "accumulation": 4, "effective_global_batch": 64, "supervised_token_loss": true, "NCCL_NVLS_ENABLE": "0", "CUDA_LAUNCH_BLOCKING": "0", "deepspeed": "scripts/zero2_h100.json"}
641
+ [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}
642
+ [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": 4, "accumulation": 4, "effective_global_batch": 64, "supervised_token_loss": true, "NCCL_NVLS_ENABLE": "0", "CUDA_LAUNCH_BLOCKING": "0", "deepspeed": "scripts/zero2_h100.json"}
643
+ [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}
644
+ [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)]
645
+ /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.
646
+ with torch.cuda.amp.autocast(dtype=dtype):
647
+ [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)]
648
+ /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.
649
+ with torch.cuda.amp.autocast(dtype=dtype):
650
+ [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": 4, "accumulation": 4, "effective_global_batch": 64, "supervised_token_loss": true, "NCCL_NVLS_ENABLE": "0", "CUDA_LAUNCH_BLOCKING": "0", "deepspeed": "scripts/zero2_h100.json"}
651
+ [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}
652
+
653
  0%| | 0/12 [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)]
654
+ /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.
655
+ with torch.cuda.amp.autocast(dtype=dtype):
656
+ [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)]
657
+ /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.
658
+ with torch.cuda.amp.autocast(dtype=dtype):
659
+ [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)]
660
+ /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.
661
+ with torch.cuda.amp.autocast(dtype=dtype):
662
+ [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)]
663
+ /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.
664
+ with torch.cuda.amp.autocast(dtype=dtype):
665
+ [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)]
666
+ /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.
667
+ with torch.cuda.amp.autocast(dtype=dtype):
668
+ [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)]
669
+ /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.
670
+ with torch.cuda.amp.autocast(dtype=dtype):
671
+ [TrainingMemory] {"rank": "2", "stage": "before_logging", "step": 1, "allocated_GiB": 24.191, "reserved_GiB": 47.305, "peak_allocated_GiB": 34.459, "peak_reserved_GiB": 47.305, "device_free_GiB": 25.647, "device_total_GiB": 79.189}
672
+ [TrainingMemory] {"rank": "3", "stage": "before_logging", "step": 1, "allocated_GiB": 24.21, "reserved_GiB": 47.088, "peak_allocated_GiB": 34.406, "peak_reserved_GiB": 47.088, "device_free_GiB": 25.866, "device_total_GiB": 79.189}
673
+ [TrainingMemory] {"rank": "0", "stage": "before_logging", "step": 1, "allocated_GiB": 24.154, "reserved_GiB": 47.693, "peak_allocated_GiB": 34.362, "peak_reserved_GiB": 47.693, "device_free_GiB": 25.259, "device_total_GiB": 79.189}
674
+
675
  8%|▊ | 1/12 [00:50<09:13, 50.31s/it][TrainingMemory] {"rank": "1", "stage": "before_logging", "step": 1, "allocated_GiB": 24.172, "reserved_GiB": 47.205, "peak_allocated_GiB": 34.4, "peak_reserved_GiB": 47.205, "device_free_GiB": 25.747, "device_total_GiB": 79.189}
676
+ [TrainingMemory] {"rank": "1", "stage": "before_logging", "step": 2, "allocated_GiB": 24.134, "reserved_GiB": 47.205, "peak_allocated_GiB": 36.441, "peak_reserved_GiB": 47.205, "device_free_GiB": 25.747, "device_total_GiB": 79.189}
677
+ [TrainingMemory] {"rank": "2", "stage": "before_logging", "step": 2, "allocated_GiB": 24.135, "reserved_GiB": 47.305, "peak_allocated_GiB": 36.493, "peak_reserved_GiB": 47.305, "device_free_GiB": 25.647, "device_total_GiB": 79.189}
678
+ [TrainingMemory] {"rank": "3", "stage": "before_logging", "step": 2, "allocated_GiB": 24.257, "reserved_GiB": 47.088, "peak_allocated_GiB": 36.486, "peak_reserved_GiB": 47.088, "device_free_GiB": 25.866, "device_total_GiB": 79.189}
679
+ [TrainingMemory] {"rank": "0", "stage": "before_logging", "step": 2, "allocated_GiB": 24.162, "reserved_GiB": 47.693, "peak_allocated_GiB": 36.192, "peak_reserved_GiB": 47.693, "device_free_GiB": 25.259, "device_total_GiB": 79.189}
680
+
681
  17%|█▋ | 2/12 [01:04<04:51, 29.11s/it][TrainingMemory] {"rank": "3", "stage": "before_logging", "step": 3, "allocated_GiB": 24.229, "reserved_GiB": 47.088, "peak_allocated_GiB": 36.386, "peak_reserved_GiB": 47.088, "device_free_GiB": 25.866, "device_total_GiB": 79.189}[TrainingMemory] {"rank": "2", "stage": "before_logging", "step": 3, "allocated_GiB": 24.257, "reserved_GiB": 47.305, "peak_allocated_GiB": 36.563, "peak_reserved_GiB": 47.305, "device_free_GiB": 25.647, "device_total_GiB": 79.189}
682
+
683
+ [TrainingMemory] {"rank": "1", "stage": "before_logging", "step": 3, "allocated_GiB": 24.2, "reserved_GiB": 47.205, "peak_allocated_GiB": 36.475, "peak_reserved_GiB": 47.205, "device_free_GiB": 25.747, "device_total_GiB": 79.189}
684
+ [TrainingMemory] {"rank": "0", "stage": "before_logging", "step": 3, "allocated_GiB": 24.163, "reserved_GiB": 47.693, "peak_allocated_GiB": 36.168, "peak_reserved_GiB": 47.693, "device_free_GiB": 25.259, "device_total_GiB": 79.189}
685
+
686
  25%|██▌ | 3/12 [01:19<03:24, 22.69s/it][TrainingMemory] {"rank": "1", "stage": "before_logging", "step": 4, "allocated_GiB": 24.173, "reserved_GiB": 47.205, "peak_allocated_GiB": 36.513, "peak_reserved_GiB": 47.205, "device_free_GiB": 25.747, "device_total_GiB": 79.189}
687
+ [TrainingMemory] {"rank": "3", "stage": "before_logging", "step": 4, "allocated_GiB": 24.153, "reserved_GiB": 47.088, "peak_allocated_GiB": 36.407, "peak_reserved_GiB": 47.088, "device_free_GiB": 25.864, "device_total_GiB": 79.189}
688
+ [TrainingMemory] {"rank": "2", "stage": "before_logging", "step": 4, "allocated_GiB": 24.181, "reserved_GiB": 47.305, "peak_allocated_GiB": 36.573, "peak_reserved_GiB": 47.305, "device_free_GiB": 25.647, "device_total_GiB": 79.189}[TrainingMemory] {"rank": "0", "stage": "before_logging", "step": 4, "allocated_GiB": 24.162, "reserved_GiB": 47.693, "peak_allocated_GiB": 36.194, "peak_reserved_GiB": 47.693, "device_free_GiB": 25.259, "device_total_GiB": 79.189}
689
+
690
+
691
  33%|███▎ | 4/12 [01:43<03:05, 23.13s/it][TrainingMemory] {"rank": "1", "stage": "before_logging", "step": 5, "allocated_GiB": 24.163, "reserved_GiB": 47.205, "peak_allocated_GiB": 36.413, "peak_reserved_GiB": 47.205, "device_free_GiB": 25.747, "device_total_GiB": 79.189}
692
+ [TrainingMemory] {"rank": "3", "stage": "before_logging", "step": 5, "allocated_GiB": 24.21, "reserved_GiB": 47.088, "peak_allocated_GiB": 36.453, "peak_reserved_GiB": 47.088, "device_free_GiB": 25.864, "device_total_GiB": 79.189}
693
+ [TrainingMemory] {"rank": "2", "stage": "before_logging", "step": 5, "allocated_GiB": 24.181, "reserved_GiB": 47.305, "peak_allocated_GiB": 36.497, "peak_reserved_GiB": 47.305, "device_free_GiB": 25.647, "device_total_GiB": 79.189}
694
+ [TrainingMemory] {"rank": "0", "stage": "before_logging", "step": 5, "allocated_GiB": 24.144, "reserved_GiB": 47.693, "peak_allocated_GiB": 36.148, "peak_reserved_GiB": 47.693, "device_free_GiB": 25.259, "device_total_GiB": 79.189}
695
+
696
  42%|████▏ | 5/12 [01:57<02:19, 19.98s/it][TrainingMemory] {"rank": "3", "stage": "before_logging", "step": 6, "allocated_GiB": 24.238, "reserved_GiB": 47.088, "peak_allocated_GiB": 36.414, "peak_reserved_GiB": 47.088, "device_free_GiB": 25.864, "device_total_GiB": 79.189}
697
+ [TrainingMemory] {"rank": "1", "stage": "before_logging", "step": 6, "allocated_GiB": 24.2, "reserved_GiB": 47.205, "peak_allocated_GiB": 36.432, "peak_reserved_GiB": 47.205, "device_free_GiB": 25.747, "device_total_GiB": 79.189}
698
+ [TrainingMemory] {"rank": "2", "stage": "before_logging", "step": 6, "allocated_GiB": 24.22, "reserved_GiB": 47.305, "peak_allocated_GiB": 36.544, "peak_reserved_GiB": 47.305, "device_free_GiB": 25.647, "device_total_GiB": 79.189}
699
+ [TrainingMemory] {"rank": "0", "stage": "before_logging", "step": 6, "allocated_GiB": 24.115, "reserved_GiB": 47.693, "peak_allocated_GiB": 36.093, "peak_reserved_GiB": 47.693, "device_free_GiB": 25.259, "device_total_GiB": 79.189}
700
+
701
  50%|█████ | 6/12 [02:12<01:49, 18.26s/it][TrainingMemory] {"rank": "3", "stage": "before_logging", "step": 7, "allocated_GiB": 24.219, "reserved_GiB": 47.088, "peak_allocated_GiB": 36.453, "peak_reserved_GiB": 47.088, "device_free_GiB": 25.864, "device_total_GiB": 79.189}
702
+ [TrainingMemory] {"rank": "0", "stage": "before_logging", "step": 7, "allocated_GiB": 24.143, "reserved_GiB": 47.693, "peak_allocated_GiB": 36.179, "peak_reserved_GiB": 47.693, "device_free_GiB": 25.259, "device_total_GiB": 79.189}
703
+
704
  58%|█████▊ | 7/12 [02:27<01:25, 17.03s/it][TrainingMemory] {"rank": "2", "stage": "before_logging", "step": 7, "allocated_GiB": 24.238, "reserved_GiB": 47.305, "peak_allocated_GiB": 36.63, "peak_reserved_GiB": 47.305, "device_free_GiB": 25.647, "device_total_GiB": 79.189}
705
+ [TrainingMemory] {"rank": "1", "stage": "before_logging", "step": 7, "allocated_GiB": 24.21, "reserved_GiB": 47.205, "peak_allocated_GiB": 36.628, "peak_reserved_GiB": 47.205, "device_free_GiB": 25.747, "device_total_GiB": 79.189}
706
+ [TrainingMemory] {"rank": "1", "stage": "before_logging", "step": 8, "allocated_GiB": 24.219, "reserved_GiB": 47.205, "peak_allocated_GiB": 36.505, "peak_reserved_GiB": 47.205, "device_free_GiB": 25.747, "device_total_GiB": 79.189}
707
+ [TrainingMemory] {"rank": "3", "stage": "before_logging", "step": 8, "allocated_GiB": 24.2, "reserved_GiB": 47.088, "peak_allocated_GiB": 36.411, "peak_reserved_GiB": 47.088, "device_free_GiB": 25.864, "device_total_GiB": 79.189}
708
+ [TrainingMemory] {"rank": "0", "stage": "before_logging", "step": 8, "allocated_GiB": 24.153, "reserved_GiB": 47.693, "peak_allocated_GiB": 36.137, "peak_reserved_GiB": 47.693, "device_free_GiB": 25.259, "device_total_GiB": 79.189}
709
+
710
  67%|██████▋ | 8/12 [02:42<01:06, 16.57s/it][TrainingMemory] {"rank": "2", "stage": "before_logging", "step": 8, "allocated_GiB": 24.191, "reserved_GiB": 47.305, "peak_allocated_GiB": 36.538, "peak_reserved_GiB": 47.305, "device_free_GiB": 25.647, "device_total_GiB": 79.189}
711
+ [TrainingMemory] {"rank": "2", "stage": "before_logging", "step": 9, "allocated_GiB": 24.229, "reserved_GiB": 47.307, "peak_allocated_GiB": 36.633, "peak_reserved_GiB": 47.307, "device_free_GiB": 25.645, "device_total_GiB": 79.189}
712
+ [TrainingMemory] {"rank": "1", "stage": "before_logging", "step": 9, "allocated_GiB": 24.219, "reserved_GiB": 47.205, "peak_allocated_GiB": 36.538, "peak_reserved_GiB": 47.205, "device_free_GiB": 25.747, "device_total_GiB": 79.189}
713
+ [TrainingMemory] {"rank": "0", "stage": "before_logging", "step": 9, "allocated_GiB": 24.181, "reserved_GiB": 47.693, "peak_allocated_GiB": 36.251, "peak_reserved_GiB": 47.693, "device_free_GiB": 25.259, "device_total_GiB": 79.189}
714
+
715
  75%|███████▌ | 9/12 [02:56<00:47, 15.75s/it][TrainingMemory] {"rank": "3", "stage": "before_logging", "step": 9, "allocated_GiB": 24.248, "reserved_GiB": 47.088, "peak_allocated_GiB": 36.51, "peak_reserved_GiB": 47.088, "device_free_GiB": 25.864, "device_total_GiB": 79.189}
716
+ [TrainingMemory] {"rank": "0", "stage": "before_logging", "step": 10, "allocated_GiB": 24.154, "reserved_GiB": 47.693, "peak_allocated_GiB": 36.114, "peak_reserved_GiB": 47.693, "device_free_GiB": 25.259, "device_total_GiB": 79.189}
717
+
718
  83%|████████▎ | 10/12 [03:12<00:31, 15.69s/it][TrainingMemory] {"rank": "3", "stage": "before_logging", "step": 10, "allocated_GiB": 24.21, "reserved_GiB": 47.088, "peak_allocated_GiB": 36.379, "peak_reserved_GiB": 47.088, "device_free_GiB": 25.864, "device_total_GiB": 79.189}
719
+ [TrainingMemory] {"rank": "2", "stage": "before_logging", "step": 10, "allocated_GiB": 24.229, "reserved_GiB": 47.307, "peak_allocated_GiB": 36.553, "peak_reserved_GiB": 47.307, "device_free_GiB": 25.645, "device_total_GiB": 79.189}
720
+ [TrainingMemory] {"rank": "1", "stage": "before_logging", "step": 10, "allocated_GiB": 24.219, "reserved_GiB": 47.205, "peak_allocated_GiB": 36.472, "peak_reserved_GiB": 47.205, "device_free_GiB": 25.747, "device_total_GiB": 79.189}
721
+ [TrainingMemory] {"rank": "1", "stage": "after_logging", "step": 10, "allocated_GiB": 24.219, "reserved_GiB": 47.205, "peak_allocated_GiB": 36.472, "peak_reserved_GiB": 47.205, "device_free_GiB": 25.274, "device_total_GiB": 79.189}
722
+ [TrainingMemory] {"rank": "2", "stage": "after_logging", "step": 10, "allocated_GiB": 24.229, "reserved_GiB": 47.307, "peak_allocated_GiB": 36.553, "peak_reserved_GiB": 47.307, "device_free_GiB": 25.173, "device_total_GiB": 79.189}
723
+ [TrainingMemory] {"rank": "0", "stage": "after_logging", "step": 10, "allocated_GiB": 24.154, "reserved_GiB": 47.693, "peak_allocated_GiB": 36.114, "peak_reserved_GiB": 47.693, "device_free_GiB": 24.786, "device_total_GiB": 79.189}
724
+ [TrainingMemory] {"rank": "3", "stage": "after_logging", "step": 10, "allocated_GiB": 24.21, "reserved_GiB": 47.088, "peak_allocated_GiB": 36.379, "peak_reserved_GiB": 47.088, "device_free_GiB": 25.391, "device_total_GiB": 79.189}
725
+
726
 
727
+
728
  83%|████████▎ | 10/12 [03:12<00:31, 15.69s/it][TrainingMemory] {"rank": "1", "stage": "before_logging", "step": 11, "allocated_GiB": 24.162, "reserved_GiB": 47.205, "peak_allocated_GiB": 36.394, "peak_reserved_GiB": 47.205, "device_free_GiB": 25.274, "device_total_GiB": 79.189}
729
+ [TrainingMemory] {"rank": "0", "stage": "before_logging", "step": 11, "allocated_GiB": 24.106, "reserved_GiB": 47.693, "peak_allocated_GiB": 36.107, "peak_reserved_GiB": 47.693, "device_free_GiB": 24.786, "device_total_GiB": 79.189}
730
+
731
  92%|█████████▏| 11/12 [03:27<00:15, 15.68s/it][TrainingMemory] {"rank": "3", "stage": "before_logging", "step": 11, "allocated_GiB": 24.153, "reserved_GiB": 47.088, "peak_allocated_GiB": 36.366, "peak_reserved_GiB": 47.088, "device_free_GiB": 25.391, "device_total_GiB": 79.189}
732
+ [TrainingMemory] {"rank": "2", "stage": "before_logging", "step": 11, "allocated_GiB": 24.162, "reserved_GiB": 47.307, "peak_allocated_GiB": 36.52, "peak_reserved_GiB": 47.307, "device_free_GiB": 25.173, "device_total_GiB": 79.189}
733
+ [TrainingMemory] {"rank": "1", "stage": "before_logging", "step": 12, "allocated_GiB": 24.191, "reserved_GiB": 47.207, "peak_allocated_GiB": 36.482, "peak_reserved_GiB": 47.207, "device_free_GiB": 25.272, "device_total_GiB": 79.189}
734
+ [TrainingMemory] {"rank": "0", "stage": "before_logging", "step": 12, "allocated_GiB": 24.192, "reserved_GiB": 47.695, "peak_allocated_GiB": 36.257, "peak_reserved_GiB": 47.695, "device_free_GiB": 24.784, "device_total_GiB": 79.189}
735
+
736
+ [TrainingMemory] {"rank": "2", "stage": "before_logging", "step": 12, "allocated_GiB": 24.172, "reserved_GiB": 47.307, "peak_allocated_GiB": 36.534, "peak_reserved_GiB": 47.307, "device_free_GiB": 25.173, "device_total_GiB": 79.189}
737
+
738
+
739
+
740
+ [TrainingMemory] {"rank": "3", "stage": "after_logging", "step": 12, "allocated_GiB": 22.888, "reserved_GiB": 47.088, "peak_allocated_GiB": 36.341, "peak_reserved_GiB": 47.088, "device_free_GiB": 25.391, "device_total_GiB": 79.189}
741
+ [TrainingMemory] {"rank": "2", "stage": "after_logging", "step": 12, "allocated_GiB": 22.888, "reserved_GiB": 47.307, "peak_allocated_GiB": 36.534, "peak_reserved_GiB": 47.307, "device_free_GiB": 25.173, "device_total_GiB": 79.189}
742
+ [TrainingMemory] {"rank": "0", "stage": "after_logging", "step": 12, "allocated_GiB": 22.888, "reserved_GiB": 47.695, "peak_allocated_GiB": 36.257, "peak_reserved_GiB": 47.695, "device_free_GiB": 24.784, "device_total_GiB": 79.189}
743
+ [TrainingMemory] {"rank": "1", "stage": "after_logging", "step": 12, "allocated_GiB": 22.888, "reserved_GiB": 47.207, "peak_allocated_GiB": 36.482, "peak_reserved_GiB": 47.207, "device_free_GiB": 25.272, "device_total_GiB": 79.189}
744
+
745
 
746
+
747
+
748
+ [rank0]:[W918 01:07:13.717497989 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())