Hugging Face's logo

AI-4-Health
/
HPP-finetuned-Meta-Llama-3-8B-Instruct

PEFT
Safetensors
trl
sft
Generated from Trainer
Model card Files Community
shin00001 commited on
Commit
98ea3fe
•
1 Parent(s): e0444cd

Upload 7 files

Browse files
Files changed (3) hide show
  1. README.md +19 -24
  2. adapter_model.safetensors +2 -2
  3. training_args.bin +0 -0
README.md CHANGED
@@ -18,7 +18,7 @@ should probably proofread and complete it, then remove this comment. -->
18
 
19
  This model is a fine-tuned version of [meta-llama/Meta-Llama-3-8B-Instruct](https://huggingface.co/meta-llama/Meta-Llama-3-8B-Instruct) on an unknown dataset.
20
  It achieves the following results on the evaluation set:
21
- - Loss: 3.1279
22
 
23
  ## Model description
24
 
@@ -46,33 +46,28 @@ The following hyperparameters were used during training:
46
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
47
  - lr_scheduler_type: cosine
48
  - lr_scheduler_warmup_ratio: 0.05
49
- - num_epochs: 70
50
  - mixed_precision_training: Native AMP
51
 
52
  ### Training results
53
 
54
- | Training Loss | Epoch | Step | Validation Loss |
55
- |:-------------:|:-------:|:----:|:---------------:|
56
- | 1.3621 | 3.4964 | 420 | 1.4379 |
57
- | 1.1954 | 6.9927 | 840 | 1.4208 |
58
- | 0.8507 | 10.4891 | 1260 | 1.5712 |
59
- | 0.789 | 13.9854 | 1680 | 1.6759 |
60
- | 0.5388 | 17.4818 | 2100 | 1.9153 |
61
- | 0.4013 | 20.9781 | 2520 | 2.0319 |
62
- | 0.2933 | 24.4745 | 2940 | 2.2094 |
63
- | 0.207 | 27.9709 | 3360 | 2.3547 |
64
- | 0.1604 | 31.4672 | 3780 | 2.5483 |
65
- | 0.1154 | 34.9636 | 4200 | 2.5953 |
66
- | 0.0982 | 38.4599 | 4620 | 2.7355 |
67
- | 0.0954 | 41.9563 | 5040 | 2.8220 |
68
- | 0.0677 | 45.4527 | 5460 | 2.8909 |
69
- | 0.0613 | 48.9490 | 5880 | 2.9654 |
70
- | 0.0482 | 52.4454 | 6300 | 3.0125 |
71
- | 0.0415 | 55.9417 | 6720 | 3.0390 |
72
- | 0.0477 | 59.4381 | 7140 | 3.0992 |
73
- | 0.0412 | 62.9344 | 7560 | 3.1126 |
74
- | 0.0327 | 66.4308 | 7980 | 3.1262 |
75
- | 0.0391 | 69.9272 | 8400 | 3.1279 |
76
 
77
 
78
  ### Framework versions
 
18
 
19
  This model is a fine-tuned version of [meta-llama/Meta-Llama-3-8B-Instruct](https://huggingface.co/meta-llama/Meta-Llama-3-8B-Instruct) on an unknown dataset.
20
  It achieves the following results on the evaluation set:
21
+ - Loss: 1.6596
22
 
23
  ## Model description
24
 
 
46
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
47
  - lr_scheduler_type: cosine
48
  - lr_scheduler_warmup_ratio: 0.05
49
+ - num_epochs: 20
50
  - mixed_precision_training: Native AMP
51
 
52
  ### Training results
53
 
54
+ | Training Loss | Epoch | Step | Validation Loss |
55
+ |:-------------:|:------:|:----:|:---------------:|
56
+ | 1.7073 | 0.4162 | 50 | 1.5993 |
57
+ | 1.4004 | 0.8325 | 100 | 1.4527 |
58
+ | 1.3051 | 1.2487 | 150 | 1.4122 |
59
+ | 1.2396 | 1.6649 | 200 | 1.3871 |
60
+ | 1.2044 | 2.0812 | 250 | 1.3906 |
61
+ | 1.1019 | 2.4974 | 300 | 1.3775 |
62
+ | 1.2682 | 2.9136 | 350 | 1.3649 |
63
+ | 1.1681 | 3.3299 | 400 | 1.4233 |
64
+ | 1.1343 | 3.7461 | 450 | 1.4160 |
65
+ | 0.7987 | 4.1623 | 500 | 1.4964 |
66
+ | 0.8663 | 4.5786 | 550 | 1.5011 |
67
+ | 0.7473 | 4.9948 | 600 | 1.4845 |
68
+ | 0.7386 | 5.4110 | 650 | 1.5706 |
69
+ | 0.61 | 5.8273 | 700 | 1.5695 |
70
+ | 0.4689 | 6.2435 | 750 | 1.6596 |
 
 
 
 
 
71
 
72
 
73
  ### Framework versions
adapter_model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:c8f5f623b5b1c86000a3a2a6f3b8bc4c7b09b9f4a1d783d2a0e5f90af49fffc9
3
- size 37774528
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4b76654e3600fee393722e7c47549703279d66c347d34487b9ed9141ea1f4ada
3
+ size 75514264
training_args.bin CHANGED
Binary files a/training_args.bin and b/training_args.bin differ