chansung commited on
Commit
edd93d9
1 Parent(s): 005f04e

Model save

Browse files
Files changed (4) hide show
  1. README.md +78 -0
  2. all_results.json +9 -0
  3. train_results.json +9 -0
  4. trainer_state.json +0 -0
README.md ADDED
@@ -0,0 +1,78 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ base_model: meta-llama/Meta-Llama-3.1-8B
3
+ datasets:
4
+ - generator
5
+ library_name: peft
6
+ license: llama3.1
7
+ tags:
8
+ - trl
9
+ - sft
10
+ - generated_from_trainer
11
+ model-index:
12
+ - name: llama3.1-8b-closedqa-gpt4o-100k
13
+ results: []
14
+ ---
15
+
16
+ <!-- This model card has been generated automatically according to the information the Trainer had access to. You
17
+ should probably proofread and complete it, then remove this comment. -->
18
+
19
+ # llama3.1-8b-closedqa-gpt4o-100k
20
+
21
+ This model is a fine-tuned version of [meta-llama/Meta-Llama-3.1-8B](https://huggingface.co/meta-llama/Meta-Llama-3.1-8B) on the generator dataset.
22
+ It achieves the following results on the evaluation set:
23
+ - Loss: 3.8424
24
+
25
+ ## Model description
26
+
27
+ More information needed
28
+
29
+ ## Intended uses & limitations
30
+
31
+ More information needed
32
+
33
+ ## Training and evaluation data
34
+
35
+ More information needed
36
+
37
+ ## Training procedure
38
+
39
+ ### Training hyperparameters
40
+
41
+ The following hyperparameters were used during training:
42
+ - learning_rate: 0.0002
43
+ - train_batch_size: 4
44
+ - eval_batch_size: 2
45
+ - seed: 42
46
+ - distributed_type: multi-GPU
47
+ - num_devices: 4
48
+ - gradient_accumulation_steps: 2
49
+ - total_train_batch_size: 32
50
+ - total_eval_batch_size: 8
51
+ - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
52
+ - lr_scheduler_type: cosine
53
+ - lr_scheduler_warmup_ratio: 0.1
54
+ - num_epochs: 10
55
+
56
+ ### Training results
57
+
58
+ | Training Loss | Epoch | Step | Validation Loss |
59
+ |:-------------:|:------:|:----:|:---------------:|
60
+ | 0.9422 | 0.9991 | 582 | 1.9762 |
61
+ | 0.8939 | 2.0 | 1165 | 2.0232 |
62
+ | 0.8255 | 2.9991 | 1747 | 2.1086 |
63
+ | 0.7584 | 4.0 | 2330 | 2.2541 |
64
+ | 0.6928 | 4.9991 | 2912 | 2.4424 |
65
+ | 0.6102 | 6.0 | 3495 | 2.7089 |
66
+ | 0.5466 | 6.9991 | 4077 | 3.0554 |
67
+ | 0.5038 | 8.0 | 4660 | 3.4053 |
68
+ | 0.4624 | 8.9991 | 5242 | 3.6952 |
69
+ | 0.454 | 9.9914 | 5820 | 3.8424 |
70
+
71
+
72
+ ### Framework versions
73
+
74
+ - PEFT 0.12.0
75
+ - Transformers 4.44.0
76
+ - Pytorch 2.4.0+cu121
77
+ - Datasets 2.20.0
78
+ - Tokenizers 0.19.1
all_results.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 9.991416309012875,
3
+ "total_flos": 8.683561975386472e+18,
4
+ "train_loss": 0.6765515476977293,
5
+ "train_runtime": 24420.5896,
6
+ "train_samples": 111440,
7
+ "train_samples_per_second": 7.63,
8
+ "train_steps_per_second": 0.238
9
+ }
train_results.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 9.991416309012875,
3
+ "total_flos": 8.683561975386472e+18,
4
+ "train_loss": 0.6765515476977293,
5
+ "train_runtime": 24420.5896,
6
+ "train_samples": 111440,
7
+ "train_samples_per_second": 7.63,
8
+ "train_steps_per_second": 0.238
9
+ }
trainer_state.json ADDED
The diff for this file is too large to render. See raw diff