Update README.md
Browse files
README.md
CHANGED
|
@@ -1,12 +1,12 @@
|
|
| 1 |
---
|
| 2 |
license: apache-2.0
|
| 3 |
base_model: Qwen/Qwen2-1.5B
|
| 4 |
-
tags:
|
| 5 |
-
- generated_from_trainer
|
| 6 |
metrics:
|
| 7 |
- accuracy
|
| 8 |
datasets:
|
| 9 |
- BEE-spoke-data/stepbasin-books
|
|
|
|
|
|
|
| 10 |
---
|
| 11 |
|
| 12 |
[<img src="https://raw.githubusercontent.com/wandb/assets/main/wandb-github-badge-28.svg" alt="Visualize in Weights & Biases" width="200" height="32"/>](https://wandb.ai/pszemraj/long-generation-tests/runs/ethp25f9)
|
|
@@ -15,52 +15,9 @@ datasets:
|
|
| 15 |
> [!IMPORTANT]
|
| 16 |
> this was finetuned at 16384 context length
|
| 17 |
|
| 18 |
-
This model is a fine-tuned version of [Qwen/Qwen2-1.5B](https://huggingface.co/Qwen/Qwen2-1.5B) on
|
|
|
|
| 19 |
It achieves the following results on the evaluation set:
|
| 20 |
- Loss: 2.8110
|
| 21 |
- Accuracy: 0.4298
|
| 22 |
-
- Num Input Tokens Seen: 44040192
|
| 23 |
-
|
| 24 |
-
## Model description
|
| 25 |
-
|
| 26 |
-
More information needed
|
| 27 |
-
|
| 28 |
-
## Intended uses & limitations
|
| 29 |
-
|
| 30 |
-
More information needed
|
| 31 |
-
|
| 32 |
-
## Training and evaluation data
|
| 33 |
-
|
| 34 |
-
More information needed
|
| 35 |
-
|
| 36 |
-
## Training procedure
|
| 37 |
-
|
| 38 |
-
### Training hyperparameters
|
| 39 |
-
|
| 40 |
-
The following hyperparameters were used during training:
|
| 41 |
-
- learning_rate: 3e-05
|
| 42 |
-
- train_batch_size: 1
|
| 43 |
-
- eval_batch_size: 1
|
| 44 |
-
- seed: 80085
|
| 45 |
-
- gradient_accumulation_steps: 32
|
| 46 |
-
- total_train_batch_size: 32
|
| 47 |
-
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 48 |
-
- lr_scheduler_type: cosine
|
| 49 |
-
- lr_scheduler_warmup_ratio: 0.05
|
| 50 |
-
- num_epochs: 3.0
|
| 51 |
-
|
| 52 |
-
### Training results
|
| 53 |
-
|
| 54 |
-
| Training Loss | Epoch | Step | Validation Loss | Accuracy | Input Tokens Seen |
|
| 55 |
-
|:-------------:|:------:|:----:|:---------------:|:--------:|:-----------------:|
|
| 56 |
-
| 2.7792 | 0.9967 | 28 | 2.8183 | 0.4287 | 14729216 |
|
| 57 |
-
| 2.6971 | 1.9933 | 56 | 2.8112 | 0.4297 | 29458432 |
|
| 58 |
-
| 2.7116 | 2.9900 | 84 | 2.8110 | 0.4298 | 44040192 |
|
| 59 |
-
|
| 60 |
-
|
| 61 |
-
### Framework versions
|
| 62 |
-
|
| 63 |
-
- Transformers 4.42.4
|
| 64 |
-
- Pytorch 2.3.1+cu121
|
| 65 |
-
- Datasets 2.20.0
|
| 66 |
-
- Tokenizers 0.19.1
|
|
|
|
| 1 |
---
|
| 2 |
license: apache-2.0
|
| 3 |
base_model: Qwen/Qwen2-1.5B
|
|
|
|
|
|
|
| 4 |
metrics:
|
| 5 |
- accuracy
|
| 6 |
datasets:
|
| 7 |
- BEE-spoke-data/stepbasin-books
|
| 8 |
+
language:
|
| 9 |
+
- en
|
| 10 |
---
|
| 11 |
|
| 12 |
[<img src="https://raw.githubusercontent.com/wandb/assets/main/wandb-github-badge-28.svg" alt="Visualize in Weights & Biases" width="200" height="32"/>](https://wandb.ai/pszemraj/long-generation-tests/runs/ethp25f9)
|
|
|
|
| 15 |
> [!IMPORTANT]
|
| 16 |
> this was finetuned at 16384 context length
|
| 17 |
|
| 18 |
+
This model is a fine-tuned version of [Qwen/Qwen2-1.5B](https://huggingface.co/Qwen/Qwen2-1.5B) on https://github.com/stepbasin/books/tree/master/books
|
| 19 |
+
|
| 20 |
It achieves the following results on the evaluation set:
|
| 21 |
- Loss: 2.8110
|
| 22 |
- Accuracy: 0.4298
|
| 23 |
+
- Num Input Tokens Seen: 44040192
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|