Migara Amarasinghe
commited on
Commit
•
f7e512d
1
Parent(s):
ca681c2
End of training
Browse files
README.md
CHANGED
@@ -5,6 +5,8 @@ tags:
|
|
5 |
- trl
|
6 |
- sft
|
7 |
- generated_from_trainer
|
|
|
|
|
8 |
base_model: google/gemma-2b
|
9 |
model-index:
|
10 |
- name: Gemma2B-LORAfied
|
@@ -16,7 +18,9 @@ should probably proofread and complete it, then remove this comment. -->
|
|
16 |
|
17 |
# Gemma2B-LORAfied
|
18 |
|
19 |
-
This model is a fine-tuned version of [google/gemma-2b](https://huggingface.co/google/gemma-2b) on
|
|
|
|
|
20 |
|
21 |
## Model description
|
22 |
|
@@ -46,6 +50,17 @@ The following hyperparameters were used during training:
|
|
46 |
- lr_scheduler_warmup_ratio: 0.05
|
47 |
- training_steps: 593
|
48 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
49 |
### Framework versions
|
50 |
|
51 |
- PEFT 0.10.0
|
|
|
5 |
- trl
|
6 |
- sft
|
7 |
- generated_from_trainer
|
8 |
+
datasets:
|
9 |
+
- generator
|
10 |
base_model: google/gemma-2b
|
11 |
model-index:
|
12 |
- name: Gemma2B-LORAfied
|
|
|
18 |
|
19 |
# Gemma2B-LORAfied
|
20 |
|
21 |
+
This model is a fine-tuned version of [google/gemma-2b](https://huggingface.co/google/gemma-2b) on the generator dataset.
|
22 |
+
It achieves the following results on the evaluation set:
|
23 |
+
- Loss: 2.1460
|
24 |
|
25 |
## Model description
|
26 |
|
|
|
50 |
- lr_scheduler_warmup_ratio: 0.05
|
51 |
- training_steps: 593
|
52 |
|
53 |
+
### Training results
|
54 |
+
|
55 |
+
| Training Loss | Epoch | Step | Validation Loss |
|
56 |
+
|:-------------:|:-----:|:----:|:---------------:|
|
57 |
+
| 2.8443 | 0.82 | 100 | 2.5332 |
|
58 |
+
| 2.4577 | 1.64 | 200 | 2.3103 |
|
59 |
+
| 2.275 | 2.46 | 300 | 2.2143 |
|
60 |
+
| 2.2331 | 3.28 | 400 | 2.1686 |
|
61 |
+
| 2.1737 | 4.1 | 500 | 2.1460 |
|
62 |
+
|
63 |
+
|
64 |
### Framework versions
|
65 |
|
66 |
- PEFT 0.10.0
|