Hugging Face's logo

prithivMLmods
/
RSI-CB256-07

Image Classification
Transformers
Safetensors
English
siglip
Remote Sensing Instruments
RSI
Location
Model card Files
xet
Community
prithivMLmods commited on
Commit
d317380
·
verified ·
1 Parent(s): b80b8a9

Upload folder using huggingface_hub

Browse files
Files changed (20) hide show
  1. checkpoint-1548/config.json +56 -0
  2. checkpoint-1548/model.safetensors +3 -0
  3. checkpoint-1548/optimizer.pt +3 -0
  4. checkpoint-1548/preprocessor_config.json +24 -0
  5. checkpoint-1548/rng_state.pth +3 -0
  6. checkpoint-1548/scheduler.pt +3 -0
  7. checkpoint-1548/trainer_state.json +75 -0
  8. checkpoint-1548/training_args.bin +3 -0
  9. checkpoint-774/config.json +56 -0
  10. checkpoint-774/model.safetensors +3 -0
  11. checkpoint-774/optimizer.pt +3 -0
  12. checkpoint-774/preprocessor_config.json +24 -0
  13. checkpoint-774/rng_state.pth +3 -0
  14. checkpoint-774/scheduler.pt +3 -0
  15. checkpoint-774/trainer_state.json +51 -0
  16. checkpoint-774/training_args.bin +3 -0
  17. config.json +56 -0
  18. model.safetensors +3 -0
  19. preprocessor_config.json +24 -0
  20. training_args.bin +3 -0
checkpoint-1548/config.json ADDED
@@ -0,0 +1,56 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "SiglipForImageClassification"
4
+ ],
5
+ "id2label": {
6
+ "0": "transportation",
7
+ "1": "other objects",
8
+ "2": "woodland",
9
+ "3": "water area",
10
+ "4": "other land",
11
+ "5": "cultivated land",
12
+ "6": "construction land"
13
+ },
14
+ "initializer_factor": 1.0,
15
+ "label2id": {
16
+ "construction land": 6,
17
+ "cultivated land": 5,
18
+ "other land": 4,
19
+ "other objects": 1,
20
+ "transportation": 0,
21
+ "water area": 3,
22
+ "woodland": 2
23
+ },
24
+ "model_type": "siglip",
25
+ "problem_type": "single_label_classification",
26
+ "text_config": {
27
+ "attention_dropout": 0.0,
28
+ "hidden_act": "gelu_pytorch_tanh",
29
+ "hidden_size": 768,
30
+ "intermediate_size": 3072,
31
+ "layer_norm_eps": 1e-06,
32
+ "max_position_embeddings": 64,
33
+ "model_type": "siglip_text_model",
34
+ "num_attention_heads": 12,
35
+ "num_hidden_layers": 12,
36
+ "projection_size": 768,
37
+ "torch_dtype": "float32",
38
+ "vocab_size": 256000
39
+ },
40
+ "torch_dtype": "float32",
41
+ "transformers_version": "4.50.0",
42
+ "vision_config": {
43
+ "attention_dropout": 0.0,
44
+ "hidden_act": "gelu_pytorch_tanh",
45
+ "hidden_size": 768,
46
+ "image_size": 224,
47
+ "intermediate_size": 3072,
48
+ "layer_norm_eps": 1e-06,
49
+ "model_type": "siglip_vision_model",
50
+ "num_attention_heads": 12,
51
+ "num_channels": 3,
52
+ "num_hidden_layers": 12,
53
+ "patch_size": 16,
54
+ "torch_dtype": "float32"
55
+ }
56
+ }
checkpoint-1548/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:202b550fc4b6b936fe44928b7c5fd111ef128ba5731c6879be2abf672091312b
3
+ size 371583372
checkpoint-1548/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c59e970cd9ec686d0ae58df32a986623a82128e8193bae27d8e0bad83735ebfb
3
+ size 686586490
checkpoint-1548/preprocessor_config.json ADDED
@@ -0,0 +1,24 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "do_convert_rgb": null,
3
+ "do_normalize": true,
4
+ "do_rescale": true,
5
+ "do_resize": true,
6
+ "image_mean": [
7
+ 0.5,
8
+ 0.5,
9
+ 0.5
10
+ ],
11
+ "image_processor_type": "SiglipImageProcessor",
12
+ "image_std": [
13
+ 0.5,
14
+ 0.5,
15
+ 0.5
16
+ ],
17
+ "processor_class": "SiglipProcessor",
18
+ "resample": 2,
19
+ "rescale_factor": 0.00392156862745098,
20
+ "size": {
21
+ "height": 224,
22
+ "width": 224
23
+ }
24
+ }
checkpoint-1548/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a127e431fd8b5f93a3bf02626092859ebe2bfdd882e2b2e537bbd13addc9cec2
3
+ size 14244
checkpoint-1548/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:cb050df0950ff8e921910b930efd06056f7447f89133e0edcb782e0a9335c2d8
3
+ size 1064
checkpoint-1548/trainer_state.json ADDED
@@ -0,0 +1,75 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_global_step": 1548,
3
+ "best_metric": 0.028825150802731514,
4
+ "best_model_checkpoint": "siglip2-finetune-full/checkpoint-1548",
5
+ "epoch": 2.0,
6
+ "eval_steps": 500,
7
+ "global_step": 1548,
8
+ "is_hyper_param_search": false,
9
+ "is_local_process_zero": true,
10
+ "is_world_process_zero": true,
11
+ "log_history": [
12
+ {
13
+ "epoch": 0.6459948320413437,
14
+ "grad_norm": 1.310089111328125,
15
+ "learning_rate": 0.00013991989319092122,
16
+ "loss": 0.5537,
17
+ "step": 500
18
+ },
19
+ {
20
+ "epoch": 1.0,
21
+ "eval_accuracy": 0.9111407443326464,
22
+ "eval_loss": 0.28554126620292664,
23
+ "eval_model_preparation_time": 0.0022,
24
+ "eval_runtime": 320.8872,
25
+ "eval_samples_per_second": 77.121,
26
+ "eval_steps_per_second": 9.642,
27
+ "step": 774
28
+ },
29
+ {
30
+ "epoch": 1.2919896640826873,
31
+ "grad_norm": 1.1912801265716553,
32
+ "learning_rate": 7.316421895861149e-05,
33
+ "loss": 0.1901,
34
+ "step": 1000
35
+ },
36
+ {
37
+ "epoch": 1.937984496124031,
38
+ "grad_norm": 2.1367239952087402,
39
+ "learning_rate": 6.408544726301736e-06,
40
+ "loss": 0.0761,
41
+ "step": 1500
42
+ },
43
+ {
44
+ "epoch": 2.0,
45
+ "eval_accuracy": 0.9912312603547905,
46
+ "eval_loss": 0.028825150802731514,
47
+ "eval_model_preparation_time": 0.0022,
48
+ "eval_runtime": 319.9983,
49
+ "eval_samples_per_second": 77.335,
50
+ "eval_steps_per_second": 9.669,
51
+ "step": 1548
52
+ }
53
+ ],
54
+ "logging_steps": 500,
55
+ "max_steps": 1548,
56
+ "num_input_tokens_seen": 0,
57
+ "num_train_epochs": 2,
58
+ "save_steps": 500,
59
+ "stateful_callbacks": {
60
+ "TrainerControl": {
61
+ "args": {
62
+ "should_epoch_stop": false,
63
+ "should_evaluate": false,
64
+ "should_log": false,
65
+ "should_save": true,
66
+ "should_training_stop": true
67
+ },
68
+ "attributes": {}
69
+ }
70
+ },
71
+ "total_flos": 4.1455663870514504e+18,
72
+ "train_batch_size": 32,
73
+ "trial_name": null,
74
+ "trial_params": null
75
+ }
checkpoint-1548/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:12dd34e788f5eb534c4b6ac8f57ace245fefa2a536daa2fe035c88d0385d96e8
3
+ size 5304
checkpoint-774/config.json ADDED
@@ -0,0 +1,56 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "SiglipForImageClassification"
4
+ ],
5
+ "id2label": {
6
+ "0": "transportation",
7
+ "1": "other objects",
8
+ "2": "woodland",
9
+ "3": "water area",
10
+ "4": "other land",
11
+ "5": "cultivated land",
12
+ "6": "construction land"
13
+ },
14
+ "initializer_factor": 1.0,
15
+ "label2id": {
16
+ "construction land": 6,
17
+ "cultivated land": 5,
18
+ "other land": 4,
19
+ "other objects": 1,
20
+ "transportation": 0,
21
+ "water area": 3,
22
+ "woodland": 2
23
+ },
24
+ "model_type": "siglip",
25
+ "problem_type": "single_label_classification",
26
+ "text_config": {
27
+ "attention_dropout": 0.0,
28
+ "hidden_act": "gelu_pytorch_tanh",
29
+ "hidden_size": 768,
30
+ "intermediate_size": 3072,
31
+ "layer_norm_eps": 1e-06,
32
+ "max_position_embeddings": 64,
33
+ "model_type": "siglip_text_model",
34
+ "num_attention_heads": 12,
35
+ "num_hidden_layers": 12,
36
+ "projection_size": 768,
37
+ "torch_dtype": "float32",
38
+ "vocab_size": 256000
39
+ },
40
+ "torch_dtype": "float32",
41
+ "transformers_version": "4.50.0",
42
+ "vision_config": {
43
+ "attention_dropout": 0.0,
44
+ "hidden_act": "gelu_pytorch_tanh",
45
+ "hidden_size": 768,
46
+ "image_size": 224,
47
+ "intermediate_size": 3072,
48
+ "layer_norm_eps": 1e-06,
49
+ "model_type": "siglip_vision_model",
50
+ "num_attention_heads": 12,
51
+ "num_channels": 3,
52
+ "num_hidden_layers": 12,
53
+ "patch_size": 16,
54
+ "torch_dtype": "float32"
55
+ }
56
+ }
checkpoint-774/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a5c923b779de2d717ee4b2316c814b74ceb52f87d15c69192026cae02193c541
3
+ size 371583372
checkpoint-774/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:641f200d0dcfd4c226db99bdcb471ead5aa6f970d08bcbfd4c761b246f41df05
3
+ size 686586490
checkpoint-774/preprocessor_config.json ADDED
@@ -0,0 +1,24 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "do_convert_rgb": null,
3
+ "do_normalize": true,
4
+ "do_rescale": true,
5
+ "do_resize": true,
6
+ "image_mean": [
7
+ 0.5,
8
+ 0.5,
9
+ 0.5
10
+ ],
11
+ "image_processor_type": "SiglipImageProcessor",
12
+ "image_std": [
13
+ 0.5,
14
+ 0.5,
15
+ 0.5
16
+ ],
17
+ "processor_class": "SiglipProcessor",
18
+ "resample": 2,
19
+ "rescale_factor": 0.00392156862745098,
20
+ "size": {
21
+ "height": 224,
22
+ "width": 224
23
+ }
24
+ }
checkpoint-774/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c77f7ded6171386b451b70d83d116dd0614aba90de316cb058986d6802fc9826
3
+ size 14244
checkpoint-774/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:db8b0e48a825de1ef288d11a2c2c7eeb56cc8010ca0357de1b98d6e304b36c63
3
+ size 1064
checkpoint-774/trainer_state.json ADDED
@@ -0,0 +1,51 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_global_step": 774,
3
+ "best_metric": 0.28554126620292664,
4
+ "best_model_checkpoint": "siglip2-finetune-full/checkpoint-774",
5
+ "epoch": 1.0,
6
+ "eval_steps": 500,
7
+ "global_step": 774,
8
+ "is_hyper_param_search": false,
9
+ "is_local_process_zero": true,
10
+ "is_world_process_zero": true,
11
+ "log_history": [
12
+ {
13
+ "epoch": 0.6459948320413437,
14
+ "grad_norm": 1.310089111328125,
15
+ "learning_rate": 0.00013991989319092122,
16
+ "loss": 0.5537,
17
+ "step": 500
18
+ },
19
+ {
20
+ "epoch": 1.0,
21
+ "eval_accuracy": 0.9111407443326464,
22
+ "eval_loss": 0.28554126620292664,
23
+ "eval_model_preparation_time": 0.0022,
24
+ "eval_runtime": 320.8872,
25
+ "eval_samples_per_second": 77.121,
26
+ "eval_steps_per_second": 9.642,
27
+ "step": 774
28
+ }
29
+ ],
30
+ "logging_steps": 500,
31
+ "max_steps": 1548,
32
+ "num_input_tokens_seen": 0,
33
+ "num_train_epochs": 2,
34
+ "save_steps": 500,
35
+ "stateful_callbacks": {
36
+ "TrainerControl": {
37
+ "args": {
38
+ "should_epoch_stop": false,
39
+ "should_evaluate": false,
40
+ "should_log": false,
41
+ "should_save": true,
42
+ "should_training_stop": false
43
+ },
44
+ "attributes": {}
45
+ }
46
+ },
47
+ "total_flos": 2.0727831935257252e+18,
48
+ "train_batch_size": 32,
49
+ "trial_name": null,
50
+ "trial_params": null
51
+ }
checkpoint-774/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:12dd34e788f5eb534c4b6ac8f57ace245fefa2a536daa2fe035c88d0385d96e8
3
+ size 5304
config.json ADDED
@@ -0,0 +1,56 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "SiglipForImageClassification"
4
+ ],
5
+ "id2label": {
6
+ "0": "transportation",
7
+ "1": "other objects",
8
+ "2": "woodland",
9
+ "3": "water area",
10
+ "4": "other land",
11
+ "5": "cultivated land",
12
+ "6": "construction land"
13
+ },
14
+ "initializer_factor": 1.0,
15
+ "label2id": {
16
+ "construction land": 6,
17
+ "cultivated land": 5,
18
+ "other land": 4,
19
+ "other objects": 1,
20
+ "transportation": 0,
21
+ "water area": 3,
22
+ "woodland": 2
23
+ },
24
+ "model_type": "siglip",
25
+ "problem_type": "single_label_classification",
26
+ "text_config": {
27
+ "attention_dropout": 0.0,
28
+ "hidden_act": "gelu_pytorch_tanh",
29
+ "hidden_size": 768,
30
+ "intermediate_size": 3072,
31
+ "layer_norm_eps": 1e-06,
32
+ "max_position_embeddings": 64,
33
+ "model_type": "siglip_text_model",
34
+ "num_attention_heads": 12,
35
+ "num_hidden_layers": 12,
36
+ "projection_size": 768,
37
+ "torch_dtype": "float32",
38
+ "vocab_size": 256000
39
+ },
40
+ "torch_dtype": "float32",
41
+ "transformers_version": "4.50.0",
42
+ "vision_config": {
43
+ "attention_dropout": 0.0,
44
+ "hidden_act": "gelu_pytorch_tanh",
45
+ "hidden_size": 768,
46
+ "image_size": 224,
47
+ "intermediate_size": 3072,
48
+ "layer_norm_eps": 1e-06,
49
+ "model_type": "siglip_vision_model",
50
+ "num_attention_heads": 12,
51
+ "num_channels": 3,
52
+ "num_hidden_layers": 12,
53
+ "patch_size": 16,
54
+ "torch_dtype": "float32"
55
+ }
56
+ }
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:202b550fc4b6b936fe44928b7c5fd111ef128ba5731c6879be2abf672091312b
3
+ size 371583372
preprocessor_config.json ADDED
@@ -0,0 +1,24 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "do_convert_rgb": null,
3
+ "do_normalize": true,
4
+ "do_rescale": true,
5
+ "do_resize": true,
6
+ "image_mean": [
7
+ 0.5,
8
+ 0.5,
9
+ 0.5
10
+ ],
11
+ "image_processor_type": "SiglipImageProcessor",
12
+ "image_std": [
13
+ 0.5,
14
+ 0.5,
15
+ 0.5
16
+ ],
17
+ "processor_class": "SiglipProcessor",
18
+ "resample": 2,
19
+ "rescale_factor": 0.00392156862745098,
20
+ "size": {
21
+ "height": 224,
22
+ "width": 224
23
+ }
24
+ }
training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:12dd34e788f5eb534c4b6ac8f57ace245fefa2a536daa2fe035c88d0385d96e8
3
+ size 5304