SneakyLemon commited on
Commit
db47e39
1 Parent(s): 783947d

SneakyLemon/Llama3LoraCauseEffect

Browse files
README.md CHANGED
@@ -17,8 +17,8 @@ should probably proofread and complete it, then remove this comment. -->
17
 
18
  This model is a fine-tuned version of [meta-llama/Meta-Llama-3-8B](https://huggingface.co/meta-llama/Meta-Llama-3-8B) on an unknown dataset.
19
  It achieves the following results on the evaluation set:
20
- - Loss: 0.4487
21
- - F1: 0.8063
22
 
23
  ## Model description
24
 
@@ -38,40 +38,77 @@ More information needed
38
 
39
  The following hyperparameters were used during training:
40
  - learning_rate: 2e-05
41
- - train_batch_size: 128
42
  - eval_batch_size: 64
43
  - seed: 42
44
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
45
  - lr_scheduler_type: linear
46
- - lr_scheduler_warmup_steps: 35
47
- - num_epochs: 3
48
  - mixed_precision_training: Native AMP
49
 
50
  ### Training results
51
 
52
  | Training Loss | Epoch | Step | Validation Loss | F1 |
53
  |:-------------:|:------:|:----:|:---------------:|:------:|
54
- | 0.8917 | 0.1368 | 16 | 0.9228 | 0.5606 |
55
- | 0.8219 | 0.2735 | 32 | 0.7617 | 0.6112 |
56
- | 0.7154 | 0.4103 | 48 | 0.6455 | 0.6687 |
57
- | 0.6278 | 0.5470 | 64 | 0.5976 | 0.6955 |
58
- | 0.5923 | 0.6838 | 80 | 0.5443 | 0.7327 |
59
- | 0.5417 | 0.8205 | 96 | 0.5212 | 0.7479 |
60
- | 0.5094 | 0.9573 | 112 | 0.5087 | 0.7586 |
61
- | 0.4866 | 1.0940 | 128 | 0.4835 | 0.7719 |
62
- | 0.4743 | 1.2308 | 144 | 0.5172 | 0.7609 |
63
- | 0.4887 | 1.3675 | 160 | 0.4905 | 0.7718 |
64
- | 0.452 | 1.5043 | 176 | 0.4706 | 0.7817 |
65
- | 0.4592 | 1.6410 | 192 | 0.4658 | 0.7795 |
66
- | 0.4372 | 1.7778 | 208 | 0.4726 | 0.7782 |
67
- | 0.4387 | 1.9145 | 224 | 0.4769 | 0.7775 |
68
- | 0.4242 | 2.0513 | 240 | 0.4526 | 0.7929 |
69
- | 0.3881 | 2.1880 | 256 | 0.4541 | 0.7975 |
70
- | 0.4081 | 2.3248 | 272 | 0.4524 | 0.8002 |
71
- | 0.3768 | 2.4615 | 288 | 0.4609 | 0.7931 |
72
- | 0.3838 | 2.5983 | 304 | 0.4511 | 0.8037 |
73
- | 0.3888 | 2.7350 | 320 | 0.4483 | 0.8011 |
74
- | 0.3791 | 2.8718 | 336 | 0.4487 | 0.8063 |
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
75
 
76
 
77
  ### Framework versions
 
17
 
18
  This model is a fine-tuned version of [meta-llama/Meta-Llama-3-8B](https://huggingface.co/meta-llama/Meta-Llama-3-8B) on an unknown dataset.
19
  It achieves the following results on the evaluation set:
20
+ - Loss: 0.4468
21
+ - F1: 0.8129
22
 
23
  ## Model description
24
 
 
38
 
39
  The following hyperparameters were used during training:
40
  - learning_rate: 2e-05
41
+ - train_batch_size: 64
42
  - eval_batch_size: 64
43
  - seed: 42
44
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
45
  - lr_scheduler_type: linear
46
+ - lr_scheduler_warmup_steps: 70
47
+ - num_epochs: 4
48
  - mixed_precision_training: Native AMP
49
 
50
  ### Training results
51
 
52
  | Training Loss | Epoch | Step | Validation Loss | F1 |
53
  |:-------------:|:------:|:----:|:---------------:|:------:|
54
+ | 1.3435 | 0.0684 | 16 | 1.1414 | 0.5363 |
55
+ | 1.1308 | 0.1368 | 32 | 1.0700 | 0.5570 |
56
+ | 0.9965 | 0.2051 | 48 | 0.9828 | 0.5715 |
57
+ | 0.9948 | 0.2735 | 64 | 0.8699 | 0.5988 |
58
+ | 0.7915 | 0.3419 | 80 | 0.7255 | 0.6388 |
59
+ | 0.7498 | 0.4103 | 96 | 0.7282 | 0.5896 |
60
+ | 0.6487 | 0.4786 | 112 | 0.6206 | 0.6762 |
61
+ | 0.6269 | 0.5470 | 128 | 0.5863 | 0.6995 |
62
+ | 0.5854 | 0.6154 | 144 | 0.5598 | 0.7262 |
63
+ | 0.591 | 0.6838 | 160 | 0.5459 | 0.7346 |
64
+ | 0.5443 | 0.7521 | 176 | 0.5346 | 0.7422 |
65
+ | 0.5621 | 0.8205 | 192 | 0.5178 | 0.7610 |
66
+ | 0.4917 | 0.8889 | 208 | 0.5109 | 0.7698 |
67
+ | 0.5397 | 0.9573 | 224 | 0.4948 | 0.7709 |
68
+ | 0.5147 | 1.0256 | 240 | 0.4932 | 0.7681 |
69
+ | 0.4582 | 1.0940 | 256 | 0.4871 | 0.7760 |
70
+ | 0.4912 | 1.1624 | 272 | 0.4807 | 0.7822 |
71
+ | 0.4322 | 1.2308 | 288 | 0.5000 | 0.7757 |
72
+ | 0.4918 | 1.2991 | 304 | 0.4779 | 0.7850 |
73
+ | 0.4611 | 1.3675 | 320 | 0.4709 | 0.7872 |
74
+ | 0.4636 | 1.4359 | 336 | 0.4753 | 0.7804 |
75
+ | 0.4458 | 1.5043 | 352 | 0.4774 | 0.7829 |
76
+ | 0.4386 | 1.5726 | 368 | 0.4629 | 0.7983 |
77
+ | 0.4551 | 1.6410 | 384 | 0.4600 | 0.7941 |
78
+ | 0.465 | 1.7094 | 400 | 0.4549 | 0.7967 |
79
+ | 0.4501 | 1.7778 | 416 | 0.4539 | 0.7966 |
80
+ | 0.4083 | 1.8462 | 432 | 0.4633 | 0.7968 |
81
+ | 0.4259 | 1.9145 | 448 | 0.4573 | 0.7959 |
82
+ | 0.4371 | 1.9829 | 464 | 0.4506 | 0.8002 |
83
+ | 0.4189 | 2.0513 | 480 | 0.4539 | 0.7901 |
84
+ | 0.4324 | 2.1197 | 496 | 0.4698 | 0.7855 |
85
+ | 0.4053 | 2.1880 | 512 | 0.4483 | 0.8042 |
86
+ | 0.3535 | 2.2564 | 528 | 0.4515 | 0.8054 |
87
+ | 0.3723 | 2.3248 | 544 | 0.4523 | 0.8091 |
88
+ | 0.3865 | 2.3932 | 560 | 0.4634 | 0.7974 |
89
+ | 0.3789 | 2.4615 | 576 | 0.4540 | 0.8031 |
90
+ | 0.4015 | 2.5299 | 592 | 0.4434 | 0.8091 |
91
+ | 0.3528 | 2.5983 | 608 | 0.4484 | 0.8085 |
92
+ | 0.3634 | 2.6667 | 624 | 0.4440 | 0.8077 |
93
+ | 0.3533 | 2.7350 | 640 | 0.4563 | 0.8004 |
94
+ | 0.3757 | 2.8034 | 656 | 0.4498 | 0.8013 |
95
+ | 0.3831 | 2.8718 | 672 | 0.4478 | 0.7993 |
96
+ | 0.3851 | 2.9402 | 688 | 0.4411 | 0.8096 |
97
+ | 0.3672 | 3.0085 | 704 | 0.4427 | 0.8115 |
98
+ | 0.3356 | 3.0769 | 720 | 0.4492 | 0.8085 |
99
+ | 0.3316 | 3.1453 | 736 | 0.4577 | 0.8108 |
100
+ | 0.3297 | 3.2137 | 752 | 0.4546 | 0.8040 |
101
+ | 0.3569 | 3.2821 | 768 | 0.4502 | 0.8070 |
102
+ | 0.3517 | 3.3504 | 784 | 0.4473 | 0.8110 |
103
+ | 0.3173 | 3.4188 | 800 | 0.4515 | 0.8092 |
104
+ | 0.3369 | 3.4872 | 816 | 0.4536 | 0.8041 |
105
+ | 0.2986 | 3.5556 | 832 | 0.4500 | 0.8115 |
106
+ | 0.3565 | 3.6239 | 848 | 0.4480 | 0.8129 |
107
+ | 0.3417 | 3.6923 | 864 | 0.4458 | 0.8120 |
108
+ | 0.3321 | 3.7607 | 880 | 0.4477 | 0.8128 |
109
+ | 0.3324 | 3.8291 | 896 | 0.4475 | 0.8123 |
110
+ | 0.338 | 3.8974 | 912 | 0.4468 | 0.8143 |
111
+ | 0.3419 | 3.9658 | 928 | 0.4468 | 0.8129 |
112
 
113
 
114
  ### Framework versions
adapter_config.json CHANGED
@@ -10,15 +10,13 @@
10
  "layers_pattern": null,
11
  "layers_to_transform": null,
12
  "loftq_config": {},
13
- "lora_alpha": 256,
14
  "lora_dropout": 0.1,
15
  "megatron_config": null,
16
  "megatron_core": "megatron.core",
17
- "modules_to_save": [
18
- "score.weight"
19
- ],
20
  "peft_type": "LORA",
21
- "r": 128,
22
  "rank_pattern": {},
23
  "revision": null,
24
  "target_modules": [
 
10
  "layers_pattern": null,
11
  "layers_to_transform": null,
12
  "loftq_config": {},
13
+ "lora_alpha": 128,
14
  "lora_dropout": 0.1,
15
  "megatron_config": null,
16
  "megatron_core": "megatron.core",
17
+ "modules_to_save": null,
 
 
18
  "peft_type": "LORA",
19
+ "r": 64,
20
  "rank_pattern": {},
21
  "revision": null,
22
  "target_modules": [
adapter_model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:14a3571a063afc9ad13aec07b18b399c9a9222fc98b135282c2b2d62f28c598c
3
- size 1268802808
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c91b9031eed3b152fe69bf4a30373bedfdf0b187105a5c4c3b963d5119100c3b
3
+ size 1159750648
runs/Jun19_19-19-02_1cfbeb3fcc7d/events.out.tfevents.1718824742.1cfbeb3fcc7d.1573.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2b461d19a3111ed080b416e9fdff22f60e58bcb3c647b1f3bf565b0e705fe20f
3
+ size 5364
runs/Jun19_19-19-28_1cfbeb3fcc7d/events.out.tfevents.1718824769.1cfbeb3fcc7d.1573.1 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9ce071f459374637dcdde1aa71026964af43695cd1285fec191a4f161fbadfee
3
+ size 36321
training_args.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:35e5048b818c7be80a0fa915e212257550fcf5cc613bb5820fde12145936c406
3
  size 5048
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1bada85ef55fb7462258f056b2e561ab9770cb07836281699caa38436d2312cc
3
  size 5048