Model save
Browse files
README.md
CHANGED
|
@@ -20,7 +20,7 @@ should probably proofread and complete it, then remove this comment. -->
|
|
| 20 |
|
| 21 |
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 the generator dataset.
|
| 22 |
It achieves the following results on the evaluation set:
|
| 23 |
-
- Loss: 0.
|
| 24 |
|
| 25 |
## Model description
|
| 26 |
|
|
@@ -40,11 +40,11 @@ More information needed
|
|
| 40 |
|
| 41 |
The following hyperparameters were used during training:
|
| 42 |
- learning_rate: 0.0002
|
| 43 |
-
- train_batch_size:
|
| 44 |
- eval_batch_size: 8
|
| 45 |
- seed: 42
|
| 46 |
- gradient_accumulation_steps: 4
|
| 47 |
-
- total_train_batch_size:
|
| 48 |
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 49 |
- lr_scheduler_type: constant
|
| 50 |
- lr_scheduler_warmup_ratio: 0.03
|
|
@@ -55,24 +55,14 @@ The following hyperparameters were used during training:
|
|
| 55 |
|
| 56 |
| Training Loss | Epoch | Step | Validation Loss |
|
| 57 |
|:-------------:|:------:|:----:|:---------------:|
|
| 58 |
-
|
|
| 59 |
-
| 0.
|
| 60 |
-
| 0.6395 | 0.4938 | 30 | 0.6430 |
|
| 61 |
-
| 0.6148 | 0.6584 | 40 | 0.6401 |
|
| 62 |
-
| 0.6257 | 0.8230 | 50 | 0.6329 |
|
| 63 |
-
| 0.5999 | 0.9877 | 60 | 0.6305 |
|
| 64 |
-
| 0.4955 | 1.1523 | 70 | 0.6351 |
|
| 65 |
-
| 0.4613 | 1.3169 | 80 | 0.6439 |
|
| 66 |
-
| 0.473 | 1.4815 | 90 | 0.6487 |
|
| 67 |
-
| 0.4793 | 1.6461 | 100 | 0.6423 |
|
| 68 |
-
| 0.4812 | 1.8107 | 110 | 0.6449 |
|
| 69 |
-
| 0.4892 | 1.9753 | 120 | 0.6391 |
|
| 70 |
|
| 71 |
|
| 72 |
### Framework versions
|
| 73 |
|
| 74 |
- PEFT 0.11.1
|
| 75 |
-
- Transformers 4.41.
|
| 76 |
-
- Pytorch 2.
|
| 77 |
- Datasets 2.19.1
|
| 78 |
- Tokenizers 0.19.1
|
|
|
|
| 20 |
|
| 21 |
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 the generator dataset.
|
| 22 |
It achieves the following results on the evaluation set:
|
| 23 |
+
- Loss: 0.6615
|
| 24 |
|
| 25 |
## Model description
|
| 26 |
|
|
|
|
| 40 |
|
| 41 |
The following hyperparameters were used during training:
|
| 42 |
- learning_rate: 0.0002
|
| 43 |
+
- train_batch_size: 2
|
| 44 |
- eval_batch_size: 8
|
| 45 |
- seed: 42
|
| 46 |
- gradient_accumulation_steps: 4
|
| 47 |
+
- total_train_batch_size: 8
|
| 48 |
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 49 |
- lr_scheduler_type: constant
|
| 50 |
- lr_scheduler_warmup_ratio: 0.03
|
|
|
|
| 55 |
|
| 56 |
| Training Loss | Epoch | Step | Validation Loss |
|
| 57 |
|:-------------:|:------:|:----:|:---------------:|
|
| 58 |
+
| 0.6769 | 0.6723 | 20 | 0.6754 |
|
| 59 |
+
| 0.5815 | 1.3445 | 40 | 0.6615 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 60 |
|
| 61 |
|
| 62 |
### Framework versions
|
| 63 |
|
| 64 |
- PEFT 0.11.1
|
| 65 |
+
- Transformers 4.41.1
|
| 66 |
+
- Pytorch 2.3.0+cu121
|
| 67 |
- Datasets 2.19.1
|
| 68 |
- Tokenizers 0.19.1
|