Hugging Face's logo

Datasets:
Gen3DF
/
tmp

Modalities:
Text
Formats:
text
Size:
< 1K
Libraries:
Datasets
Dataset card Data Studio Files Community
1
RunyiY commited on
Commit
4d98a12
·
verified ·
1 Parent(s): 37bd193

Add files using upload-large-folder tool

Browse files
Files changed (8) hide show
  1. all_configs.yaml +111 -0
  2. best/best_loss.pth +3 -0
  3. best/best_loss.txt +194 -0
  4. checkpoints/00570.model.pth +3 -0
  5. checkpoints/00570.solver.tar +3 -0
  6. events.out.tfevents.1750714448.dgx4.plus.discoverer.bg.3074541.0 +3 -0
  7. events.out.tfevents.1750770533.dgx4.plus.discoverer.bg.3629272.0 +3 -0
  8. spatialLM_scannetpp.yaml +52 -0
all_configs.yaml ADDED
@@ -0,0 +1,111 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ BASE:
2
+ - ''
3
+ DATA:
4
+ test:
5
+ angle: &id001
6
+ - 180
7
+ - 180
8
+ - 180
9
+ batch_size: 1
10
+ codetemplate: utils/code_template.txt
11
+ depth: 5
12
+ disable: false
13
+ distort: false
14
+ filelist: data/ScanNetpp.txt
15
+ flip: &id002
16
+ - 0.0
17
+ - 0.0
18
+ - 0.0
19
+ full_depth: 2
20
+ interval: &id003
21
+ - 1
22
+ - 1
23
+ - 1
24
+ jitter: 0.0
25
+ location: /home/runyi_yang/Gen3D/dataset/ScanNetpp-SpatialLM/
26
+ name: SpatialLM
27
+ num_workers: 0
28
+ orient_normal: ''
29
+ pin_memory: true
30
+ scale: 0.0
31
+ shuffle: false
32
+ take: -1
33
+ tokenizer_name: manycore-research/SpatialLM-Llama-1B
34
+ uniform: false
35
+ train:
36
+ angle: *id001
37
+ batch_size: 2
38
+ codetemplate: utils/code_template.txt
39
+ config_path: checkpoints/SpatialLM-Llama-1B
40
+ depth: 5
41
+ disable: false
42
+ distort: false
43
+ filelist: data/ScanNetpp.txt
44
+ flip: *id002
45
+ full_depth: 2
46
+ interval: *id003
47
+ jitter: 0.0
48
+ location: /home/runyi_yang/Gen3D/dataset/ScanNetpp-SpatialLM/
49
+ name: SpatialLM
50
+ num_workers: 16
51
+ orient_normal: ''
52
+ pin_memory: true
53
+ scale: 0.0
54
+ shuffle: true
55
+ take: -1
56
+ tokenizer_name: manycore-research/SpatialLM-Llama-1B
57
+ uniform: false
58
+ LOSS:
59
+ name: ''
60
+ MODEL:
61
+ channel: 3
62
+ config_path: checkpoints/SpatialLM-Llama-1B
63
+ feature: ND
64
+ find_unused_parameters: true
65
+ model_name: SpatialLMLlama
66
+ name: ''
67
+ nempty: false
68
+ sync_bn: false
69
+ tokenizer_name: manycore-research/SpatialLM-Llama-1B
70
+ use_checkpoint: false
71
+ SOLVER:
72
+ alias: ''
73
+ best_val: min:loss
74
+ ckpt: ''
75
+ ckpt_num: 5
76
+ clip_grad: -1.0
77
+ empty_cache: 50
78
+ eval_epoch: 1
79
+ eval_step: -1
80
+ expand_ckpt: false
81
+ gamma: 0.1
82
+ gpu:
83
+ - 0
84
+ - 1
85
+ - 2
86
+ log_per_iter: 5
87
+ logdir: logs/spatialLM_scannetpp
88
+ lr: 0.0001
89
+ lr_min: 0.0001
90
+ lr_power: 0.9
91
+ lr_type: cos
92
+ max_epoch: 1500
93
+ milestones:
94
+ - 120
95
+ - 180
96
+ port: 10001
97
+ progress_bar: true
98
+ rand_seed: 0
99
+ run: train
100
+ step_size:
101
+ - 160
102
+ - 240
103
+ test_every_epoch: 1
104
+ type: adamw
105
+ use_amp: true
106
+ warmup_epoch: 20
107
+ warmup_init: 0.001
108
+ weight_decay: 0.01
109
+ zero_grad_to_none: false
110
+ SYS:
111
+ cmds: python main_spatiallm.py --config configs/runyi/spatialLM_scannetpp.yaml
best/best_loss.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:546b922e29f9ca163b01823fae8e899d3b5f1747d0d89dd4a9449371d3af6bc0
3
+ size 4989519750
best/best_loss.txt ADDED
@@ -0,0 +1,194 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ New best loss: 8.768831
2
+ New best loss: 7.910408
3
+ New best loss: 6.804373
4
+ New best loss: 6.068866
5
+ New best loss: 5.416229
6
+ New best loss: 5.053037
7
+ New best loss: 4.527080
8
+ Epoch 1 - New best epoch loss: 6.403154
9
+ Epoch 2 - New best epoch loss: 4.722594
10
+ Epoch 3 - New best epoch loss: 4.392240
11
+ Epoch 4 - New best epoch loss: 4.206342
12
+ Epoch 5 - New best epoch loss: 4.060041
13
+ Epoch 6 - New best epoch loss: 3.556629
14
+ Epoch 7 - New best epoch loss: 3.335292
15
+ Epoch 8 - New best epoch loss: 3.186932
16
+ Epoch 9 - New best epoch loss: 3.054582
17
+ Epoch 10 - New best epoch loss: 2.961526
18
+ Epoch 11 - New best epoch loss: 2.919272
19
+ Epoch 12 - New best epoch loss: 2.905780
20
+ Epoch 13 - New best epoch loss: 2.839159
21
+ Epoch 14 - New best epoch loss: 2.815196
22
+ Epoch 15 - New best epoch loss: 2.647150
23
+ Epoch 16 - New best epoch loss: 2.584679
24
+ Epoch 17 - New best epoch loss: 2.577846
25
+ Epoch 18 - New best epoch loss: 2.442734
26
+ Epoch 19 - New best epoch loss: 2.338746
27
+ Epoch 20 - New best epoch loss: 2.212231
28
+ Epoch 21 - New best epoch loss: 2.032574
29
+ Epoch 22 - New best epoch loss: 1.976113
30
+ Epoch 23 - New best epoch loss: 1.874807
31
+ Epoch 24 - New best epoch loss: 1.858073
32
+ Epoch 25 - New best epoch loss: 1.847367
33
+ Epoch 26 - New best epoch loss: 1.795330
34
+ Epoch 27 - New best epoch loss: 1.739876
35
+ Epoch 28 - New best epoch loss: 1.710324
36
+ Epoch 29 - New best epoch loss: 1.679662
37
+ Epoch 31 - New best epoch loss: 1.639029
38
+ Epoch 36 - New best epoch loss: 1.627906
39
+ Epoch 37 - New best epoch loss: 1.569179
40
+ Epoch 38 - New best epoch loss: 1.524848
41
+ Epoch 40 - New best epoch loss: 1.507860
42
+ Epoch 43 - New best epoch loss: 1.488921
43
+ Epoch 44 - New best epoch loss: 1.481632
44
+ Epoch 45 - New best epoch loss: 1.472959
45
+ Epoch 46 - New best epoch loss: 1.445820
46
+ Epoch 49 - New best epoch loss: 1.424756
47
+ Epoch 50 - New best epoch loss: 1.411044
48
+ Epoch 51 - New best epoch loss: 1.395253
49
+ Epoch 53 - New best epoch loss: 1.375894
50
+ Epoch 58 - New best epoch loss: 1.356351
51
+ Epoch 63 - New best epoch loss: 1.343870
52
+ Epoch 66 - New best epoch loss: 1.341170
53
+ Epoch 67 - New best epoch loss: 1.336127
54
+ Epoch 68 - New best epoch loss: 1.324269
55
+ Epoch 69 - New best epoch loss: 1.306407
56
+ Epoch 70 - New best epoch loss: 1.304688
57
+ Epoch 71 - New best epoch loss: 1.303475
58
+ Epoch 72 - New best epoch loss: 1.297826
59
+ Epoch 75 - New best epoch loss: 1.286101
60
+ Epoch 78 - New best epoch loss: 1.278212
61
+ Epoch 79 - New best epoch loss: 1.268077
62
+ Epoch 81 - New best epoch loss: 1.257994
63
+ Epoch 86 - New best epoch loss: 1.247838
64
+ Epoch 87 - New best epoch loss: 1.241101
65
+ Epoch 91 - New best epoch loss: 1.238606
66
+ Epoch 92 - New best epoch loss: 1.238250
67
+ Epoch 95 - New best epoch loss: 1.225926
68
+ Epoch 97 - New best epoch loss: 1.222177
69
+ Epoch 98 - New best epoch loss: 1.215219
70
+ Epoch 104 - New best epoch loss: 1.212296
71
+ Epoch 105 - New best epoch loss: 1.205576
72
+ Epoch 106 - New best epoch loss: 1.197628
73
+ Epoch 107 - New best epoch loss: 1.192027
74
+ Epoch 111 - New best epoch loss: 1.191832
75
+ Epoch 112 - New best epoch loss: 1.186528
76
+ Epoch 115 - New best epoch loss: 1.176364
77
+ Epoch 117 - New best epoch loss: 1.168163
78
+ Epoch 119 - New best epoch loss: 1.165005
79
+ Epoch 128 - New best epoch loss: 1.152950
80
+ Epoch 129 - New best epoch loss: 1.142698
81
+ Epoch 136 - New best epoch loss: 1.136441
82
+ Epoch 140 - New best epoch loss: 1.132770
83
+ Epoch 141 - New best epoch loss: 1.125236
84
+ Epoch 147 - New best epoch loss: 1.121195
85
+ Epoch 150 - New best epoch loss: 1.109871
86
+ Epoch 152 - New best epoch loss: 1.099315
87
+ Epoch 155 - New best epoch loss: 1.098736
88
+ Epoch 158 - New best epoch loss: 1.089501
89
+ Epoch 159 - New best epoch loss: 1.089129
90
+ Epoch 160 - New best epoch loss: 1.076973
91
+ Epoch 161 - New best epoch loss: 1.071318
92
+ Epoch 167 - New best epoch loss: 1.069940
93
+ Epoch 168 - New best epoch loss: 1.063515
94
+ Epoch 171 - New best epoch loss: 1.059205
95
+ Epoch 173 - New best epoch loss: 1.055928
96
+ Epoch 176 - New best epoch loss: 1.037312
97
+ Epoch 183 - New best epoch loss: 1.036732
98
+ Epoch 184 - New best epoch loss: 1.029775
99
+ Epoch 185 - New best epoch loss: 1.025906
100
+ Epoch 187 - New best epoch loss: 1.022229
101
+ Epoch 189 - New best epoch loss: 1.014456
102
+ Epoch 190 - New best epoch loss: 1.011433
103
+ Epoch 191 - New best epoch loss: 1.008803
104
+ Epoch 192 - New best epoch loss: 1.008417
105
+ Epoch 195 - New best epoch loss: 0.998204
106
+ Epoch 199 - New best epoch loss: 0.996210
107
+ Epoch 200 - New best epoch loss: 0.987192
108
+ Epoch 201 - New best epoch loss: 0.984024
109
+ Epoch 203 - New best epoch loss: 0.973925
110
+ Epoch 208 - New best epoch loss: 0.973756
111
+ Epoch 209 - New best epoch loss: 0.973136
112
+ Epoch 210 - New best epoch loss: 0.945097
113
+ Epoch 217 - New best epoch loss: 0.936184
114
+ Epoch 220 - New best epoch loss: 0.933823
115
+ Epoch 223 - New best epoch loss: 0.919811
116
+ Epoch 228 - New best epoch loss: 0.905661
117
+ Epoch 229 - New best epoch loss: 0.905203
118
+ Epoch 230 - New best epoch loss: 0.901975
119
+ Epoch 232 - New best epoch loss: 0.901890
120
+ Epoch 236 - New best epoch loss: 0.898593
121
+ Epoch 237 - New best epoch loss: 0.891758
122
+ Epoch 238 - New best epoch loss: 0.886407
123
+ Epoch 242 - New best epoch loss: 0.885633
124
+ Epoch 244 - New best epoch loss: 0.873728
125
+ Epoch 246 - New best epoch loss: 0.864800
126
+ Epoch 247 - New best epoch loss: 0.855866
127
+ Epoch 248 - New best epoch loss: 0.844989
128
+ Epoch 254 - New best epoch loss: 0.834004
129
+ Epoch 263 - New best epoch loss: 0.823090
130
+ Epoch 265 - New best epoch loss: 0.817603
131
+ Epoch 266 - New best epoch loss: 0.814879
132
+ Epoch 269 - New best epoch loss: 0.803147
133
+ Epoch 271 - New best epoch loss: 0.791444
134
+ Epoch 279 - New best epoch loss: 0.783689
135
+ Epoch 280 - New best epoch loss: 0.783106
136
+ Epoch 282 - New best epoch loss: 0.764783
137
+ Epoch 286 - New best epoch loss: 0.764176
138
+ Epoch 288 - New best epoch loss: 0.762750
139
+ Epoch 289 - New best epoch loss: 0.740439
140
+ Epoch 299 - New best epoch loss: 0.738873
141
+ Epoch 301 - New best epoch loss: 0.729100
142
+ Epoch 303 - New best epoch loss: 0.719293
143
+ Epoch 304 - New best epoch loss: 0.713386
144
+ Epoch 305 - New best epoch loss: 0.708755
145
+ Epoch 310 - New best epoch loss: 0.687669
146
+ Epoch 321 - New best epoch loss: 0.684264
147
+ Epoch 324 - New best epoch loss: 0.679451
148
+ Epoch 326 - New best epoch loss: 0.672125
149
+ Epoch 327 - New best epoch loss: 0.644865
150
+ Epoch 339 - New best epoch loss: 0.644511
151
+ Epoch 340 - New best epoch loss: 0.641004
152
+ Epoch 347 - New best epoch loss: 0.626519
153
+ Epoch 349 - New best epoch loss: 0.618889
154
+ Epoch 355 - New best epoch loss: 0.596862
155
+ Epoch 359 - New best epoch loss: 0.591170
156
+ Epoch 370 - New best epoch loss: 0.584298
157
+ Epoch 371 - New best epoch loss: 0.573251
158
+ Epoch 375 - New best epoch loss: 0.549142
159
+ Epoch 386 - New best epoch loss: 0.546385
160
+ Epoch 390 - New best epoch loss: 0.524576
161
+ Epoch 400 - New best epoch loss: 0.519864
162
+ Epoch 404 - New best epoch loss: 0.516361
163
+ Epoch 407 - New best epoch loss: 0.508423
164
+ Epoch 410 - New best epoch loss: 0.503608
165
+ Epoch 415 - New best epoch loss: 0.483310
166
+ Epoch 432 - New best epoch loss: 0.474851
167
+ Epoch 435 - New best epoch loss: 0.470614
168
+ Epoch 437 - New best epoch loss: 0.464458
169
+ Epoch 438 - New best epoch loss: 0.471630
170
+ Epoch 440 - New best epoch loss: 0.457099
171
+ Epoch 444 - New best epoch loss: 0.451216
172
+ Epoch 451 - New best epoch loss: 0.448762
173
+ Epoch 455 - New best epoch loss: 0.428194
174
+ Epoch 468 - New best epoch loss: 0.426060
175
+ Epoch 471 - New best epoch loss: 0.424199
176
+ Epoch 472 - New best epoch loss: 0.423968
177
+ Epoch 473 - New best epoch loss: 0.414455
178
+ Epoch 474 - New best epoch loss: 0.410689
179
+ Epoch 486 - New best epoch loss: 0.399772
180
+ Epoch 490 - New best epoch loss: 0.397970
181
+ Epoch 500 - New best epoch loss: 0.377831
182
+ Epoch 505 - New best epoch loss: 0.365189
183
+ Epoch 506 - New best epoch loss: 0.364544
184
+ Epoch 509 - New best epoch loss: 0.362200
185
+ Epoch 519 - New best epoch loss: 0.362183
186
+ Epoch 520 - New best epoch loss: 0.360367
187
+ Epoch 527 - New best epoch loss: 0.355875
188
+ Epoch 529 - New best epoch loss: 0.347712
189
+ Epoch 536 - New best epoch loss: 0.336020
190
+ Epoch 543 - New best epoch loss: 0.335477
191
+ Epoch 551 - New best epoch loss: 0.323956
192
+ Epoch 558 - New best epoch loss: 0.312646
193
+ Epoch 559 - New best epoch loss: 0.307570
194
+ Epoch 568 - New best epoch loss: 0.302366
checkpoints/00570.model.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:97c3daacb75baffa8ba3c1adca696ffa023dba0bc02ced8c398f303168c678d2
3
+ size 4989520210
checkpoints/00570.solver.tar ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e9b7e5eae941eca20e159a113ac9c852320cece4040790498d29a4a94a6c818e
3
+ size 5081922384
events.out.tfevents.1750714448.dgx4.plus.discoverer.bg.3074541.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a6b638201527f2273a6a3eda565a5c0f10c15092f12e73086aa0af3073036ea8
3
+ size 62635
events.out.tfevents.1750770533.dgx4.plus.discoverer.bg.3629272.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f85ee7392e9086b63a2716527d92af0a3b40b7b89f363cb4b48b332a2fc52e90
3
+ size 19240
spatialLM_scannetpp.yaml ADDED
@@ -0,0 +1,52 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ SOLVER:
2
+ gpu: 0,1,2,
3
+ run: train
4
+ logdir: logs/spatialLM_scannetpp
5
+ max_epoch: 1500
6
+ test_every_epoch: 1
7
+ log_per_iter: 5
8
+ ckpt_num: 5
9
+ expand_ckpt: False
10
+ port: 10001
11
+
12
+ # optimizer
13
+ type: adamw
14
+ weight_decay: 0.01 # default value of adamw
15
+ lr: 0.0001 # default value of adamw
16
+ rand_seed: 0
17
+ use_amp: True
18
+
19
+ # learning rate
20
+ lr_type: cos
21
+ step_size: (160,240)
22
+
23
+ DATA:
24
+ train:
25
+ name: &name SpatialLM
26
+ # data loading
27
+ location: &location /home/runyi_yang/Gen3D/dataset/ScanNetpp-SpatialLM/
28
+ filelist: &filelist data/ScanNetpp.txt
29
+ codetemplate: &codetemplate utils/code_template.txt
30
+ tokenizer_name: &tokenizer_name manycore-research/SpatialLM-Llama-1B # meta-llama/Llama-3.2-1B-Instruct
31
+ config_path: &config_path checkpoints/SpatialLM-Llama-1B
32
+ batch_size: 2
33
+ shuffle: True
34
+ num_workers: 16
35
+
36
+ test:
37
+ name: *name
38
+ # data loading
39
+ location: *location
40
+ filelist: data/ScanNetpp.txt
41
+ codetemplate: *codetemplate
42
+ tokenizer_name: *tokenizer_name
43
+ batch_size: 1
44
+ shuffle: False
45
+ num_workers: 0
46
+
47
+
48
+ MODEL:
49
+ find_unused_parameters: True
50
+ model_name: SpatialLMLlama
51
+ tokenizer_name: *tokenizer_name
52
+ config_path: *config_path