Update README.md
Browse files
README.md
CHANGED
|
@@ -1,6 +1,6 @@
|
|
| 1 |
---
|
| 2 |
library_name: peft
|
| 3 |
-
base_model: LSX-UniWue/
|
| 4 |
tags:
|
| 5 |
- trl
|
| 6 |
- sft
|
|
@@ -14,53 +14,7 @@ language:
|
|
| 14 |
- de
|
| 15 |
---
|
| 16 |
|
| 17 |
-
|
| 18 |
-
should probably proofread and complete it, then remove this comment. -->
|
| 19 |
|
| 20 |
-
|
| 21 |
-
|
| 22 |
-
This model is a fine-tuned version of [LSX-UniWue/LLaMmlein_1b](https://huggingface.co/LSX-UniWue/LLaMmlein_1b) on an unknown dataset.
|
| 23 |
-
It achieves the following results on the evaluation set:
|
| 24 |
-
- Loss: 1.6298
|
| 25 |
-
|
| 26 |
-
## Model description
|
| 27 |
-
|
| 28 |
-
More information needed
|
| 29 |
-
|
| 30 |
-
## Intended uses & limitations
|
| 31 |
-
|
| 32 |
-
More information needed
|
| 33 |
-
|
| 34 |
-
## Training and evaluation data
|
| 35 |
-
|
| 36 |
-
More information needed
|
| 37 |
-
|
| 38 |
-
## Training procedure
|
| 39 |
-
|
| 40 |
-
### Training hyperparameters
|
| 41 |
-
|
| 42 |
-
The following hyperparameters were used during training:
|
| 43 |
-
- learning_rate: 5e-05
|
| 44 |
-
- train_batch_size: 8
|
| 45 |
-
- eval_batch_size: 8
|
| 46 |
-
- seed: 42
|
| 47 |
-
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 48 |
-
- lr_scheduler_type: linear
|
| 49 |
-
- num_epochs: 3.0
|
| 50 |
-
|
| 51 |
-
### Training results
|
| 52 |
-
|
| 53 |
-
| Training Loss | Epoch | Step | Validation Loss |
|
| 54 |
-
|:-------------:|:-----:|:----:|:---------------:|
|
| 55 |
-
| 1.533 | 1.0 | 763 | 1.6431 |
|
| 56 |
-
| 1.4528 | 2.0 | 1526 | 1.6315 |
|
| 57 |
-
| 1.1119 | 3.0 | 2289 | 1.6298 |
|
| 58 |
-
|
| 59 |
-
|
| 60 |
-
### Framework versions
|
| 61 |
-
|
| 62 |
-
- PEFT 0.13.2
|
| 63 |
-
- Transformers 4.44.2
|
| 64 |
-
- Pytorch 2.4.1+cu121
|
| 65 |
-
- Datasets 3.0.1
|
| 66 |
-
- Tokenizers 0.19.1
|
|
|
|
| 1 |
---
|
| 2 |
library_name: peft
|
| 3 |
+
base_model: LSX-UniWue/LLaMmlein_1B
|
| 4 |
tags:
|
| 5 |
- trl
|
| 6 |
- sft
|
|
|
|
| 14 |
- de
|
| 15 |
---
|
| 16 |
|
| 17 |
+
# LLäMmlein 1B Chat
|
|
|
|
| 18 |
|
| 19 |
+
This is a chat adapter for the German Tinyllama 1B language model.
|
| 20 |
+
Find more details on our [page](https://www.informatik.uni-wuerzburg.de/datascience/projects/nlp/llammlein/)!
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|