alexander-hm commited on
Commit
cba43fb
1 Parent(s): 93f04f9

End of training

Browse files
Files changed (7) hide show
  1. README.md +72 -0
  2. all_results.json +12 -0
  3. completed +0 -0
  4. eval_results.json +7 -0
  5. metrics.json +1 -0
  6. train_results.json +8 -0
  7. trainer_state.json +0 -0
README.md ADDED
@@ -0,0 +1,72 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ base_model: huggyllama/llama-13b
3
+ library_name: peft
4
+ license: other
5
+ tags:
6
+ - generated_from_trainer
7
+ model-index:
8
+ - name: llama-13b_oasst1_l0.0002_32-8-8-8-8
9
+ results: []
10
+ ---
11
+
12
+ <!-- This model card has been generated automatically according to the information the Trainer had access to. You
13
+ should probably proofread and complete it, then remove this comment. -->
14
+
15
+ # llama-13b_oasst1_l0.0002_32-8-8-8-8
16
+
17
+ This model is a fine-tuned version of [huggyllama/llama-13b](https://huggingface.co/huggyllama/llama-13b) on an unknown dataset.
18
+ It achieves the following results on the evaluation set:
19
+ - Loss: 1.5115
20
+
21
+ ## Model description
22
+
23
+ More information needed
24
+
25
+ ## Intended uses & limitations
26
+
27
+ More information needed
28
+
29
+ ## Training and evaluation data
30
+
31
+ More information needed
32
+
33
+ ## Training procedure
34
+
35
+ ### Training hyperparameters
36
+
37
+ The following hyperparameters were used during training:
38
+ - learning_rate: 0.0002
39
+ - train_batch_size: 1
40
+ - eval_batch_size: 1
41
+ - seed: 0
42
+ - gradient_accumulation_steps: 16
43
+ - total_train_batch_size: 16
44
+ - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
45
+ - lr_scheduler_type: constant
46
+ - lr_scheduler_warmup_ratio: 0.03
47
+ - training_steps: 1875
48
+
49
+ ### Training results
50
+
51
+ | Training Loss | Epoch | Step | Validation Loss |
52
+ |:-------------:|:------:|:----:|:---------------:|
53
+ | 1.4264 | 0.0018 | 1 | 1.5678 |
54
+ | 1.4263 | 0.3392 | 187 | 1.2365 |
55
+ | 1.0712 | 0.6783 | 374 | 1.2313 |
56
+ | 1.2232 | 1.0175 | 561 | 1.2366 |
57
+ | 0.9766 | 1.3566 | 748 | 1.2977 |
58
+ | 0.983 | 1.6958 | 935 | 1.2653 |
59
+ | 0.5607 | 2.0349 | 1122 | 1.3461 |
60
+ | 0.7534 | 2.3741 | 1309 | 1.3852 |
61
+ | 0.5426 | 2.7132 | 1496 | 1.4808 |
62
+ | 0.2964 | 3.0524 | 1683 | 1.5780 |
63
+ | 0.3246 | 3.3915 | 1870 | 1.4927 |
64
+
65
+
66
+ ### Framework versions
67
+
68
+ - PEFT 0.12.1.dev0
69
+ - Transformers 4.45.0.dev0
70
+ - Pytorch 2.3.0+cu121
71
+ - Datasets 2.19.0
72
+ - Tokenizers 0.19.1
all_results.json ADDED
@@ -0,0 +1,12 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 3.4005894355021535,
3
+ "eval_loss": 1.511498212814331,
4
+ "eval_runtime": 331.6938,
5
+ "eval_samples_per_second": 3.015,
6
+ "eval_steps_per_second": 3.015,
7
+ "total_flos": 8.246064732932506e+17,
8
+ "train_loss": 0.8989817124764125,
9
+ "train_runtime": 128394.3068,
10
+ "train_samples_per_second": 0.234,
11
+ "train_steps_per_second": 0.015
12
+ }
completed ADDED
File without changes
eval_results.json ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 3.4005894355021535,
3
+ "eval_loss": 1.511498212814331,
4
+ "eval_runtime": 331.6938,
5
+ "eval_samples_per_second": 3.015,
6
+ "eval_steps_per_second": 3.015
7
+ }
metrics.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"run_name": "huggyllama/llama-13b_oasst1_l0.0002_32,8,8,8,8", "train_runtime": 128394.3068, "train_samples_per_second": 0.234, "train_steps_per_second": 0.015, "total_flos": 8.246064732932506e+17, "train_loss": 0.8989817124764125, "epoch": 3.4005894355021535, "eval_loss": 1.511498212814331, "eval_runtime": 331.6938, "eval_samples_per_second": 3.015, "eval_steps_per_second": 3.015}
train_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 3.4005894355021535,
3
+ "total_flos": 8.246064732932506e+17,
4
+ "train_loss": 0.8989817124764125,
5
+ "train_runtime": 128394.3068,
6
+ "train_samples_per_second": 0.234,
7
+ "train_steps_per_second": 0.015
8
+ }
trainer_state.json ADDED
The diff for this file is too large to render. See raw diff