Hugging Face's logo

kailinjiang
/
MMKE-Bench

TensorBoard
Model card Files
xet
Metrics Community
Yifan Zhong commited on
Commit
37b9a44
·
1 Parent(s): f7d7ebb

Add new file

Browse files
Files changed (27) hide show
  1. KE-Models/entity/blip2/version_0/checkpoints/model-epoch=15-valid_acc=0.6215.ckpt +3 -0
  2. KE-Models/entity/blip2/version_0/events.out.tfevents.1729486028.hgx-hyperplane06.3100839.0 +3 -0
  3. KE-Models/entity/blip2/version_0/hparams.yaml +77 -0
  4. KE-Models/entity/llava/version_0/checkpoints/model-epoch=30-valid_acc=0.7554.ckpt +3 -0
  5. KE-Models/entity/llava/version_0/events.out.tfevents.1729480997.hgx-hyperplane00.560963.0 +3 -0
  6. KE-Models/entity/llava/version_0/hparams.yaml +77 -0
  7. KE-Models/entity/minigpt4/version_0/checkpoints/model-epoch=30-valid_acc=0.7267.ckpt +3 -0
  8. KE-Models/entity/minigpt4/version_0/events.out.tfevents.1729500183.hgx-hyperplane00.839260.0 +3 -0
  9. KE-Models/entity/minigpt4/version_0/hparams.yaml +77 -0
  10. KE-Models/user/blip2/version_0/checkpoints/model-epoch=53-valid_acc=0.6875.ckpt +3 -0
  11. KE-Models/user/blip2/version_0/events.out.tfevents.1729494709.hgx-hyperplane08.4053573.0 +3 -0
  12. KE-Models/user/blip2/version_0/hparams.yaml +77 -0
  13. KE-Models/user/llava/version_0/checkpoints/model-epoch=57-valid_acc=0.8655.ckpt +3 -0
  14. KE-Models/user/llava/version_0/events.out.tfevents.1729474375.hgx-hyperplane02.1416491.0 +3 -0
  15. KE-Models/user/llava/version_0/hparams.yaml +77 -0
  16. KE-Models/user/minigpt4/version_0/checkpoints/model-epoch=38-valid_acc=0.8061.ckpt +3 -0
  17. KE-Models/user/minigpt4/version_0/events.out.tfevents.1729532566.hgx-hyperplane06.3235774.0 +3 -0
  18. KE-Models/user/minigpt4/version_0/hparams.yaml +77 -0
  19. KE-Models/visual/blip2/version_0/checkpoints/model-epoch=86-valid_acc=0.6090.ckpt +3 -0
  20. KE-Models/visual/blip2/version_0/events.out.tfevents.1729497331.hgx-hyperplane06.3134266.0 +3 -0
  21. KE-Models/visual/blip2/version_0/hparams.yaml +77 -0
  22. KE-Models/visual/llava/version_0/checkpoints/model-epoch=69-valid_acc=0.7734.ckpt +3 -0
  23. KE-Models/visual/llava/version_0/events.out.tfevents.1729478404.hgx-hyperplane00.549392.0 +3 -0
  24. KE-Models/visual/llava/version_0/hparams.yaml +77 -0
  25. KE-Models/visual/minigpt4/version_0/checkpoints/model-epoch=87-valid_acc=0.7183.ckpt +3 -0
  26. KE-Models/visual/minigpt4/version_0/events.out.tfevents.1729534393.hgx-hyperplane00.1156861.0 +3 -0
  27. KE-Models/visual/minigpt4/version_0/hparams.yaml +77 -0
KE-Models/entity/blip2/version_0/checkpoints/model-epoch=15-valid_acc=0.6215.ckpt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d8fff65d27ff47a5cbe9588614651d07910cb76afaed6e20b5f2f0c6a3373c21
3
+ size 1829697707
KE-Models/entity/blip2/version_0/events.out.tfevents.1729486028.hgx-hyperplane06.3100839.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ed07a4905de2b77a7bb84f20408e364a191052efa5df903dce8bf7aa2dd10b37
3
+ size 120494
KE-Models/entity/blip2/version_0/hparams.yaml ADDED
@@ -0,0 +1,77 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ dirpath: models/
2
+ save_top_k: 1
3
+ seed: 0
4
+ data_type: entity
5
+ train_data_path: data_json/entity_train.json
6
+ dev_data_path: data_json/entity_eval.json
7
+ batch_size: 1
8
+ lr: 0.0003
9
+ lr_alpha: 0.1
10
+ max_length: 32
11
+ weight_decay: 0.01
12
+ total_num_updates: 20000
13
+ warmup_updates: 1000
14
+ num_workers: 0
15
+ model_name: blip2
16
+ eps: 0.1
17
+ model_checkpoint: ''
18
+ margin_lp_max: 0.001
19
+ margin_lp_min: 1.0e-07
20
+ max_scale: 1
21
+ p: 2
22
+ divergences: lp
23
+ use_views: false
24
+ logger: true
25
+ checkpoint_callback: true
26
+ default_root_dir: null
27
+ gradient_clip_val: 0.0
28
+ gradient_clip_algorithm: norm
29
+ process_position: 0
30
+ num_nodes: 1
31
+ num_processes: 1
32
+ gpus: -1
33
+ auto_select_gpus: false
34
+ tpu_cores: null
35
+ log_gpu_memory: null
36
+ progress_bar_refresh_rate: null
37
+ overfit_batches: 0.0
38
+ track_grad_norm: -1
39
+ check_val_every_n_epoch: 1
40
+ fast_dev_run: false
41
+ accumulate_grad_batches: 1
42
+ max_epochs: null
43
+ min_epochs: null
44
+ max_steps: 20000
45
+ min_steps: null
46
+ max_time: null
47
+ limit_train_batches: 1.0
48
+ limit_val_batches: 1.0
49
+ limit_test_batches: 1.0
50
+ limit_predict_batches: 1.0
51
+ val_check_interval: 1.0
52
+ flush_logs_every_n_steps: 100
53
+ log_every_n_steps: 50
54
+ accelerator: null
55
+ sync_batchnorm: false
56
+ precision: 32
57
+ weights_summary: top
58
+ weights_save_path: null
59
+ num_sanity_val_steps: 2
60
+ truncated_bptt_steps: null
61
+ resume_from_checkpoint: null
62
+ profiler: null
63
+ benchmark: false
64
+ deterministic: false
65
+ reload_dataloaders_every_epoch: false
66
+ auto_lr_find: false
67
+ replace_sampler_ddp: true
68
+ terminate_on_nan: false
69
+ auto_scale_batch_size: false
70
+ prepare_data_per_node: true
71
+ plugins: null
72
+ amp_backend: native
73
+ amp_level: O2
74
+ distributed_backend: null
75
+ move_metrics_to_cpu: false
76
+ multiple_trainloader_mode: max_size_cycle
77
+ stochastic_weight_avg: false
KE-Models/entity/llava/version_0/checkpoints/model-epoch=30-valid_acc=0.7554.ckpt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:702029fe2e4567a3d88154edaa98fc4087e186017a9dcb62b406225214cb7a80
3
+ size 1920210795
KE-Models/entity/llava/version_0/events.out.tfevents.1729480997.hgx-hyperplane00.560963.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:462a237c997f797ff31bf35f5df8945e3670724b3b053afadd63dc67a42126ea
3
+ size 120494
KE-Models/entity/llava/version_0/hparams.yaml ADDED
@@ -0,0 +1,77 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ dirpath: models/
2
+ save_top_k: 1
3
+ seed: 0
4
+ data_type: entity
5
+ train_data_path: data_json/entity_train.json
6
+ dev_data_path: data_json/entity_eval.json
7
+ batch_size: 1
8
+ lr: 0.0003
9
+ lr_alpha: 0.1
10
+ max_length: 32
11
+ weight_decay: 0.01
12
+ total_num_updates: 20000
13
+ warmup_updates: 1000
14
+ num_workers: 0
15
+ model_name: llava
16
+ eps: 0.1
17
+ model_checkpoint: ''
18
+ margin_lp_max: 0.001
19
+ margin_lp_min: 1.0e-07
20
+ max_scale: 1
21
+ p: 2
22
+ divergences: lp
23
+ use_views: false
24
+ logger: true
25
+ checkpoint_callback: true
26
+ default_root_dir: null
27
+ gradient_clip_val: 0.0
28
+ gradient_clip_algorithm: norm
29
+ process_position: 0
30
+ num_nodes: 1
31
+ num_processes: 1
32
+ gpus: -1
33
+ auto_select_gpus: false
34
+ tpu_cores: null
35
+ log_gpu_memory: null
36
+ progress_bar_refresh_rate: null
37
+ overfit_batches: 0.0
38
+ track_grad_norm: -1
39
+ check_val_every_n_epoch: 1
40
+ fast_dev_run: false
41
+ accumulate_grad_batches: 1
42
+ max_epochs: null
43
+ min_epochs: null
44
+ max_steps: 20000
45
+ min_steps: null
46
+ max_time: null
47
+ limit_train_batches: 1.0
48
+ limit_val_batches: 1.0
49
+ limit_test_batches: 1.0
50
+ limit_predict_batches: 1.0
51
+ val_check_interval: 1.0
52
+ flush_logs_every_n_steps: 100
53
+ log_every_n_steps: 50
54
+ accelerator: null
55
+ sync_batchnorm: false
56
+ precision: 32
57
+ weights_summary: top
58
+ weights_save_path: null
59
+ num_sanity_val_steps: 2
60
+ truncated_bptt_steps: null
61
+ resume_from_checkpoint: null
62
+ profiler: null
63
+ benchmark: false
64
+ deterministic: false
65
+ reload_dataloaders_every_epoch: false
66
+ auto_lr_find: false
67
+ replace_sampler_ddp: true
68
+ terminate_on_nan: false
69
+ auto_scale_batch_size: false
70
+ prepare_data_per_node: true
71
+ plugins: null
72
+ amp_backend: native
73
+ amp_level: O2
74
+ distributed_backend: null
75
+ move_metrics_to_cpu: false
76
+ multiple_trainloader_mode: max_size_cycle
77
+ stochastic_weight_avg: false
KE-Models/entity/minigpt4/version_0/checkpoints/model-epoch=30-valid_acc=0.7267.ckpt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:dab7a757659e0bb8d6258f639bb9850e9a6a3a8d900c18ad737d37f9b7c20fe4
3
+ size 1920260395
KE-Models/entity/minigpt4/version_0/events.out.tfevents.1729500183.hgx-hyperplane00.839260.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9ac3f0527315173154c1168009f7ea5c66c81e0735fc0361399703fc1aac9c19
3
+ size 120497
KE-Models/entity/minigpt4/version_0/hparams.yaml ADDED
@@ -0,0 +1,77 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ dirpath: models/
2
+ save_top_k: 1
3
+ seed: 0
4
+ data_type: entity
5
+ train_data_path: data_json/entity_train.json
6
+ dev_data_path: data_json/entity_eval.json
7
+ batch_size: 1
8
+ lr: 0.0003
9
+ lr_alpha: 0.1
10
+ max_length: 32
11
+ weight_decay: 0.01
12
+ total_num_updates: 20000
13
+ warmup_updates: 1000
14
+ num_workers: 0
15
+ model_name: minigpt4
16
+ eps: 0.1
17
+ model_checkpoint: ''
18
+ margin_lp_max: 0.001
19
+ margin_lp_min: 1.0e-07
20
+ max_scale: 1
21
+ p: 2
22
+ divergences: lp
23
+ use_views: false
24
+ logger: true
25
+ checkpoint_callback: true
26
+ default_root_dir: null
27
+ gradient_clip_val: 0.0
28
+ gradient_clip_algorithm: norm
29
+ process_position: 0
30
+ num_nodes: 1
31
+ num_processes: 1
32
+ gpus: -1
33
+ auto_select_gpus: false
34
+ tpu_cores: null
35
+ log_gpu_memory: null
36
+ progress_bar_refresh_rate: null
37
+ overfit_batches: 0.0
38
+ track_grad_norm: -1
39
+ check_val_every_n_epoch: 1
40
+ fast_dev_run: false
41
+ accumulate_grad_batches: 1
42
+ max_epochs: null
43
+ min_epochs: null
44
+ max_steps: 20000
45
+ min_steps: null
46
+ max_time: null
47
+ limit_train_batches: 1.0
48
+ limit_val_batches: 1.0
49
+ limit_test_batches: 1.0
50
+ limit_predict_batches: 1.0
51
+ val_check_interval: 1.0
52
+ flush_logs_every_n_steps: 100
53
+ log_every_n_steps: 50
54
+ accelerator: null
55
+ sync_batchnorm: false
56
+ precision: 32
57
+ weights_summary: top
58
+ weights_save_path: null
59
+ num_sanity_val_steps: 2
60
+ truncated_bptt_steps: null
61
+ resume_from_checkpoint: null
62
+ profiler: null
63
+ benchmark: false
64
+ deterministic: false
65
+ reload_dataloaders_every_epoch: false
66
+ auto_lr_find: false
67
+ replace_sampler_ddp: true
68
+ terminate_on_nan: false
69
+ auto_scale_batch_size: false
70
+ prepare_data_per_node: true
71
+ plugins: null
72
+ amp_backend: native
73
+ amp_level: O2
74
+ distributed_backend: null
75
+ move_metrics_to_cpu: false
76
+ multiple_trainloader_mode: max_size_cycle
77
+ stochastic_weight_avg: false
KE-Models/user/blip2/version_0/checkpoints/model-epoch=53-valid_acc=0.6875.ckpt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7a77c968bc1db79f7a3208a1d5d855a5252eb39e01c36d988d0739431a436441
3
+ size 1829697707
KE-Models/user/blip2/version_0/events.out.tfevents.1729494709.hgx-hyperplane08.4053573.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9b40d06e436a1667df14456fd843325518b08880a0b8555a5822c3602f7e0625
3
+ size 124563
KE-Models/user/blip2/version_0/hparams.yaml ADDED
@@ -0,0 +1,77 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ dirpath: models/
2
+ save_top_k: 1
3
+ seed: 0
4
+ data_type: user
5
+ train_data_path: data_json/user_train.json
6
+ dev_data_path: data_json/user_eval.json
7
+ batch_size: 1
8
+ lr: 0.0003
9
+ lr_alpha: 0.1
10
+ max_length: 32
11
+ weight_decay: 0.01
12
+ total_num_updates: 20000
13
+ warmup_updates: 1000
14
+ num_workers: 0
15
+ model_name: blip2
16
+ eps: 0.1
17
+ model_checkpoint: ''
18
+ margin_lp_max: 0.001
19
+ margin_lp_min: 1.0e-07
20
+ max_scale: 1
21
+ p: 2
22
+ divergences: lp
23
+ use_views: false
24
+ logger: true
25
+ checkpoint_callback: true
26
+ default_root_dir: null
27
+ gradient_clip_val: 0.0
28
+ gradient_clip_algorithm: norm
29
+ process_position: 0
30
+ num_nodes: 1
31
+ num_processes: 1
32
+ gpus: -1
33
+ auto_select_gpus: false
34
+ tpu_cores: null
35
+ log_gpu_memory: null
36
+ progress_bar_refresh_rate: null
37
+ overfit_batches: 0.0
38
+ track_grad_norm: -1
39
+ check_val_every_n_epoch: 1
40
+ fast_dev_run: false
41
+ accumulate_grad_batches: 1
42
+ max_epochs: null
43
+ min_epochs: null
44
+ max_steps: 20000
45
+ min_steps: null
46
+ max_time: null
47
+ limit_train_batches: 1.0
48
+ limit_val_batches: 1.0
49
+ limit_test_batches: 1.0
50
+ limit_predict_batches: 1.0
51
+ val_check_interval: 1.0
52
+ flush_logs_every_n_steps: 100
53
+ log_every_n_steps: 50
54
+ accelerator: null
55
+ sync_batchnorm: false
56
+ precision: 32
57
+ weights_summary: top
58
+ weights_save_path: null
59
+ num_sanity_val_steps: 2
60
+ truncated_bptt_steps: null
61
+ resume_from_checkpoint: null
62
+ profiler: null
63
+ benchmark: false
64
+ deterministic: false
65
+ reload_dataloaders_every_epoch: false
66
+ auto_lr_find: false
67
+ replace_sampler_ddp: true
68
+ terminate_on_nan: false
69
+ auto_scale_batch_size: false
70
+ prepare_data_per_node: true
71
+ plugins: null
72
+ amp_backend: native
73
+ amp_level: O2
74
+ distributed_backend: null
75
+ move_metrics_to_cpu: false
76
+ multiple_trainloader_mode: max_size_cycle
77
+ stochastic_weight_avg: false
KE-Models/user/llava/version_0/checkpoints/model-epoch=57-valid_acc=0.8655.ckpt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c78459d35ca57a9b5ce32c9ebf0617dc34b540ea42d8d07b9116ff4f9bd6874a
3
+ size 1920210795
KE-Models/user/llava/version_0/events.out.tfevents.1729474375.hgx-hyperplane02.1416491.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4447cf0d45acde009e6dcbd93d7f8d3c72211f438e5a438a044c390f2bc0b327
3
+ size 124563
KE-Models/user/llava/version_0/hparams.yaml ADDED
@@ -0,0 +1,77 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ dirpath: models/
2
+ save_top_k: 1
3
+ seed: 0
4
+ data_type: user
5
+ train_data_path: data_json/user_train.json
6
+ dev_data_path: data_json/user_eval.json
7
+ batch_size: 1
8
+ lr: 0.0003
9
+ lr_alpha: 0.1
10
+ max_length: 32
11
+ weight_decay: 0.01
12
+ total_num_updates: 20000
13
+ warmup_updates: 1000
14
+ num_workers: 0
15
+ model_name: llava
16
+ eps: 0.1
17
+ model_checkpoint: ''
18
+ margin_lp_max: 0.001
19
+ margin_lp_min: 1.0e-07
20
+ max_scale: 1
21
+ p: 2
22
+ divergences: lp
23
+ use_views: false
24
+ logger: true
25
+ checkpoint_callback: true
26
+ default_root_dir: null
27
+ gradient_clip_val: 0.0
28
+ gradient_clip_algorithm: norm
29
+ process_position: 0
30
+ num_nodes: 1
31
+ num_processes: 1
32
+ gpus: -1
33
+ auto_select_gpus: false
34
+ tpu_cores: null
35
+ log_gpu_memory: null
36
+ progress_bar_refresh_rate: null
37
+ overfit_batches: 0.0
38
+ track_grad_norm: -1
39
+ check_val_every_n_epoch: 1
40
+ fast_dev_run: false
41
+ accumulate_grad_batches: 1
42
+ max_epochs: null
43
+ min_epochs: null
44
+ max_steps: 20000
45
+ min_steps: null
46
+ max_time: null
47
+ limit_train_batches: 1.0
48
+ limit_val_batches: 1.0
49
+ limit_test_batches: 1.0
50
+ limit_predict_batches: 1.0
51
+ val_check_interval: 1.0
52
+ flush_logs_every_n_steps: 100
53
+ log_every_n_steps: 50
54
+ accelerator: null
55
+ sync_batchnorm: false
56
+ precision: 32
57
+ weights_summary: top
58
+ weights_save_path: null
59
+ num_sanity_val_steps: 2
60
+ truncated_bptt_steps: null
61
+ resume_from_checkpoint: null
62
+ profiler: null
63
+ benchmark: false
64
+ deterministic: false
65
+ reload_dataloaders_every_epoch: false
66
+ auto_lr_find: false
67
+ replace_sampler_ddp: true
68
+ terminate_on_nan: false
69
+ auto_scale_batch_size: false
70
+ prepare_data_per_node: true
71
+ plugins: null
72
+ amp_backend: native
73
+ amp_level: O2
74
+ distributed_backend: null
75
+ move_metrics_to_cpu: false
76
+ multiple_trainloader_mode: max_size_cycle
77
+ stochastic_weight_avg: false
KE-Models/user/minigpt4/version_0/checkpoints/model-epoch=38-valid_acc=0.8061.ckpt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2e3ecb365bd0f218286b4cb1a9fcbc060d804003754207a665488911e4e6c649
3
+ size 1920260395
KE-Models/user/minigpt4/version_0/events.out.tfevents.1729532566.hgx-hyperplane06.3235774.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f6c73d59ea609dd20278ce67b37ed47ab0f292e9fc303a2b1d1aa650cb7cb331
3
+ size 124566
KE-Models/user/minigpt4/version_0/hparams.yaml ADDED
@@ -0,0 +1,77 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ dirpath: models/
2
+ save_top_k: 1
3
+ seed: 0
4
+ data_type: user
5
+ train_data_path: data_json/user_train.json
6
+ dev_data_path: data_json/user_eval.json
7
+ batch_size: 1
8
+ lr: 0.0003
9
+ lr_alpha: 0.1
10
+ max_length: 32
11
+ weight_decay: 0.01
12
+ total_num_updates: 20000
13
+ warmup_updates: 1000
14
+ num_workers: 0
15
+ model_name: minigpt4
16
+ eps: 0.1
17
+ model_checkpoint: ''
18
+ margin_lp_max: 0.001
19
+ margin_lp_min: 1.0e-07
20
+ max_scale: 1
21
+ p: 2
22
+ divergences: lp
23
+ use_views: false
24
+ logger: true
25
+ checkpoint_callback: true
26
+ default_root_dir: null
27
+ gradient_clip_val: 0.0
28
+ gradient_clip_algorithm: norm
29
+ process_position: 0
30
+ num_nodes: 1
31
+ num_processes: 1
32
+ gpus: -1
33
+ auto_select_gpus: false
34
+ tpu_cores: null
35
+ log_gpu_memory: null
36
+ progress_bar_refresh_rate: null
37
+ overfit_batches: 0.0
38
+ track_grad_norm: -1
39
+ check_val_every_n_epoch: 1
40
+ fast_dev_run: false
41
+ accumulate_grad_batches: 1
42
+ max_epochs: null
43
+ min_epochs: null
44
+ max_steps: 20000
45
+ min_steps: null
46
+ max_time: null
47
+ limit_train_batches: 1.0
48
+ limit_val_batches: 1.0
49
+ limit_test_batches: 1.0
50
+ limit_predict_batches: 1.0
51
+ val_check_interval: 1.0
52
+ flush_logs_every_n_steps: 100
53
+ log_every_n_steps: 50
54
+ accelerator: null
55
+ sync_batchnorm: false
56
+ precision: 32
57
+ weights_summary: top
58
+ weights_save_path: null
59
+ num_sanity_val_steps: 2
60
+ truncated_bptt_steps: null
61
+ resume_from_checkpoint: null
62
+ profiler: null
63
+ benchmark: false
64
+ deterministic: false
65
+ reload_dataloaders_every_epoch: false
66
+ auto_lr_find: false
67
+ replace_sampler_ddp: true
68
+ terminate_on_nan: false
69
+ auto_scale_batch_size: false
70
+ prepare_data_per_node: true
71
+ plugins: null
72
+ amp_backend: native
73
+ amp_level: O2
74
+ distributed_backend: null
75
+ move_metrics_to_cpu: false
76
+ multiple_trainloader_mode: max_size_cycle
77
+ stochastic_weight_avg: false
KE-Models/visual/blip2/version_0/checkpoints/model-epoch=86-valid_acc=0.6090.ckpt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ec02f42712eb56782ac7f86d01adc95c54631ad8df64669e57e53c224aba23c4
3
+ size 1829697707
KE-Models/visual/blip2/version_0/events.out.tfevents.1729497331.hgx-hyperplane06.3134266.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b49f6cda9175f187d71a2182fdbc4fdb2369c278d24a4f6622d877ec0fe1645e
3
+ size 129207
KE-Models/visual/blip2/version_0/hparams.yaml ADDED
@@ -0,0 +1,77 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ dirpath: models/
2
+ save_top_k: 1
3
+ seed: 0
4
+ data_type: visual
5
+ train_data_path: data_json/visual_train.json
6
+ dev_data_path: data_json/visual_eval.json
7
+ batch_size: 1
8
+ lr: 0.0003
9
+ lr_alpha: 0.1
10
+ max_length: 32
11
+ weight_decay: 0.01
12
+ total_num_updates: 20000
13
+ warmup_updates: 1000
14
+ num_workers: 0
15
+ model_name: blip2
16
+ eps: 0.1
17
+ model_checkpoint: ''
18
+ margin_lp_max: 0.001
19
+ margin_lp_min: 1.0e-07
20
+ max_scale: 1
21
+ p: 2
22
+ divergences: lp
23
+ use_views: false
24
+ logger: true
25
+ checkpoint_callback: true
26
+ default_root_dir: null
27
+ gradient_clip_val: 0.0
28
+ gradient_clip_algorithm: norm
29
+ process_position: 0
30
+ num_nodes: 1
31
+ num_processes: 1
32
+ gpus: -1
33
+ auto_select_gpus: false
34
+ tpu_cores: null
35
+ log_gpu_memory: null
36
+ progress_bar_refresh_rate: null
37
+ overfit_batches: 0.0
38
+ track_grad_norm: -1
39
+ check_val_every_n_epoch: 1
40
+ fast_dev_run: false
41
+ accumulate_grad_batches: 1
42
+ max_epochs: null
43
+ min_epochs: null
44
+ max_steps: 20000
45
+ min_steps: null
46
+ max_time: null
47
+ limit_train_batches: 1.0
48
+ limit_val_batches: 1.0
49
+ limit_test_batches: 1.0
50
+ limit_predict_batches: 1.0
51
+ val_check_interval: 1.0
52
+ flush_logs_every_n_steps: 100
53
+ log_every_n_steps: 50
54
+ accelerator: null
55
+ sync_batchnorm: false
56
+ precision: 32
57
+ weights_summary: top
58
+ weights_save_path: null
59
+ num_sanity_val_steps: 2
60
+ truncated_bptt_steps: null
61
+ resume_from_checkpoint: null
62
+ profiler: null
63
+ benchmark: false
64
+ deterministic: false
65
+ reload_dataloaders_every_epoch: false
66
+ auto_lr_find: false
67
+ replace_sampler_ddp: true
68
+ terminate_on_nan: false
69
+ auto_scale_batch_size: false
70
+ prepare_data_per_node: true
71
+ plugins: null
72
+ amp_backend: native
73
+ amp_level: O2
74
+ distributed_backend: null
75
+ move_metrics_to_cpu: false
76
+ multiple_trainloader_mode: max_size_cycle
77
+ stochastic_weight_avg: false
KE-Models/visual/llava/version_0/checkpoints/model-epoch=69-valid_acc=0.7734.ckpt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:987d41f5ad191b38f8d6e2df55a032bdc3ba77b115ddd4900c1710dfda77655a
3
+ size 1920210795
KE-Models/visual/llava/version_0/events.out.tfevents.1729478404.hgx-hyperplane00.549392.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b4d31400181169cf50bbba820ad1322bc60e99414d608c148ef205ca9f438da8
3
+ size 129207
KE-Models/visual/llava/version_0/hparams.yaml ADDED
@@ -0,0 +1,77 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ dirpath: models/
2
+ save_top_k: 1
3
+ seed: 0
4
+ data_type: visual
5
+ train_data_path: data_json/visual_train.json
6
+ dev_data_path: data_json/visual_eval.json
7
+ batch_size: 1
8
+ lr: 0.0003
9
+ lr_alpha: 0.1
10
+ max_length: 32
11
+ weight_decay: 0.01
12
+ total_num_updates: 20000
13
+ warmup_updates: 1000
14
+ num_workers: 0
15
+ model_name: llava
16
+ eps: 0.1
17
+ model_checkpoint: ''
18
+ margin_lp_max: 0.001
19
+ margin_lp_min: 1.0e-07
20
+ max_scale: 1
21
+ p: 2
22
+ divergences: lp
23
+ use_views: false
24
+ logger: true
25
+ checkpoint_callback: true
26
+ default_root_dir: null
27
+ gradient_clip_val: 0.0
28
+ gradient_clip_algorithm: norm
29
+ process_position: 0
30
+ num_nodes: 1
31
+ num_processes: 1
32
+ gpus: -1
33
+ auto_select_gpus: false
34
+ tpu_cores: null
35
+ log_gpu_memory: null
36
+ progress_bar_refresh_rate: null
37
+ overfit_batches: 0.0
38
+ track_grad_norm: -1
39
+ check_val_every_n_epoch: 1
40
+ fast_dev_run: false
41
+ accumulate_grad_batches: 1
42
+ max_epochs: null
43
+ min_epochs: null
44
+ max_steps: 20000
45
+ min_steps: null
46
+ max_time: null
47
+ limit_train_batches: 1.0
48
+ limit_val_batches: 1.0
49
+ limit_test_batches: 1.0
50
+ limit_predict_batches: 1.0
51
+ val_check_interval: 1.0
52
+ flush_logs_every_n_steps: 100
53
+ log_every_n_steps: 50
54
+ accelerator: null
55
+ sync_batchnorm: false
56
+ precision: 32
57
+ weights_summary: top
58
+ weights_save_path: null
59
+ num_sanity_val_steps: 2
60
+ truncated_bptt_steps: null
61
+ resume_from_checkpoint: null
62
+ profiler: null
63
+ benchmark: false
64
+ deterministic: false
65
+ reload_dataloaders_every_epoch: false
66
+ auto_lr_find: false
67
+ replace_sampler_ddp: true
68
+ terminate_on_nan: false
69
+ auto_scale_batch_size: false
70
+ prepare_data_per_node: true
71
+ plugins: null
72
+ amp_backend: native
73
+ amp_level: O2
74
+ distributed_backend: null
75
+ move_metrics_to_cpu: false
76
+ multiple_trainloader_mode: max_size_cycle
77
+ stochastic_weight_avg: false
KE-Models/visual/minigpt4/version_0/checkpoints/model-epoch=87-valid_acc=0.7183.ckpt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8833c6494197463b9b34e56fd12dcf6e91298734e1230e00645f5e3c09eaf17d
3
+ size 1920260395
KE-Models/visual/minigpt4/version_0/events.out.tfevents.1729534393.hgx-hyperplane00.1156861.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4787e62f61885e3f6008d41b49221dba2330c00b3d1755878850a95856239d6d
3
+ size 129210
KE-Models/visual/minigpt4/version_0/hparams.yaml ADDED
@@ -0,0 +1,77 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ dirpath: models/
2
+ save_top_k: 1
3
+ seed: 0
4
+ data_type: visual
5
+ train_data_path: data_json/visual_train.json
6
+ dev_data_path: data_json/visual_eval.json
7
+ batch_size: 1
8
+ lr: 0.0003
9
+ lr_alpha: 0.1
10
+ max_length: 32
11
+ weight_decay: 0.01
12
+ total_num_updates: 20000
13
+ warmup_updates: 1000
14
+ num_workers: 0
15
+ model_name: minigpt4
16
+ eps: 0.1
17
+ model_checkpoint: ''
18
+ margin_lp_max: 0.001
19
+ margin_lp_min: 1.0e-07
20
+ max_scale: 1
21
+ p: 2
22
+ divergences: lp
23
+ use_views: false
24
+ logger: true
25
+ checkpoint_callback: true
26
+ default_root_dir: null
27
+ gradient_clip_val: 0.0
28
+ gradient_clip_algorithm: norm
29
+ process_position: 0
30
+ num_nodes: 1
31
+ num_processes: 1
32
+ gpus: -1
33
+ auto_select_gpus: false
34
+ tpu_cores: null
35
+ log_gpu_memory: null
36
+ progress_bar_refresh_rate: null
37
+ overfit_batches: 0.0
38
+ track_grad_norm: -1
39
+ check_val_every_n_epoch: 1
40
+ fast_dev_run: false
41
+ accumulate_grad_batches: 1
42
+ max_epochs: null
43
+ min_epochs: null
44
+ max_steps: 20000
45
+ min_steps: null
46
+ max_time: null
47
+ limit_train_batches: 1.0
48
+ limit_val_batches: 1.0
49
+ limit_test_batches: 1.0
50
+ limit_predict_batches: 1.0
51
+ val_check_interval: 1.0
52
+ flush_logs_every_n_steps: 100
53
+ log_every_n_steps: 50
54
+ accelerator: null
55
+ sync_batchnorm: false
56
+ precision: 32
57
+ weights_summary: top
58
+ weights_save_path: null
59
+ num_sanity_val_steps: 2
60
+ truncated_bptt_steps: null
61
+ resume_from_checkpoint: null
62
+ profiler: null
63
+ benchmark: false
64
+ deterministic: false
65
+ reload_dataloaders_every_epoch: false
66
+ auto_lr_find: false
67
+ replace_sampler_ddp: true
68
+ terminate_on_nan: false
69
+ auto_scale_batch_size: false
70
+ prepare_data_per_node: true
71
+ plugins: null
72
+ amp_backend: native
73
+ amp_level: O2
74
+ distributed_backend: null
75
+ move_metrics_to_cpu: false
76
+ multiple_trainloader_mode: max_size_cycle
77
+ stochastic_weight_avg: false