File size: 1,763 Bytes
9180178 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 |
---
license: other
library_name: peft
tags:
- llama-factory
- lora
- generated_from_trainer
base_model: /ML-A100/team/mm/eamon/self_instruction/models/Qwen1_5_14B
model-index:
- name: qwen_14B_d1_iter2_model
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# qwen_14B_d1_iter2_model
This model is a fine-tuned version of [/ML-A100/team/mm/eamon/self_instruction/models/Qwen1_5_14B](https://huggingface.co//ML-A100/team/mm/eamon/self_instruction/models/Qwen1_5_14B) on the qwen_14B_d1_iter2_model dataset.
It achieves the following results on the evaluation set:
- Loss: 0.5391
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 5e-05
- train_batch_size: 1
- eval_batch_size: 1
- seed: 42
- distributed_type: multi-GPU
- num_devices: 8
- gradient_accumulation_steps: 2
- total_train_batch_size: 16
- total_eval_batch_size: 8
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- lr_scheduler_warmup_steps: 20
- num_epochs: 2.0
### Training results
| Training Loss | Epoch | Step | Validation Loss |
|:-------------:|:------:|:----:|:---------------:|
| 0.5923 | 0.5917 | 100 | 0.5549 |
| 0.485 | 1.1834 | 200 | 0.5391 |
| 0.4405 | 1.7751 | 300 | 0.5399 |
### Framework versions
- PEFT 0.10.0
- Transformers 4.41.2
- Pytorch 2.1.2+cu121
- Datasets 2.18.0
- Tokenizers 0.19.1 |