metadata
base_model: /data/modelscope/qwen/Qwen2-7B-Instruct
library_name: peft
license: other
tags:
- llama-factory
- lora
- generated_from_trainer
model-index:
- name: sft
results: []
sft
This model is a fine-tuned version of /data/modelscope/qwen/Qwen2-7B-Instruct on the llm-complex-reasoning-train-qwen2-72b-instruct-correct and the Infinity-Instruct-0625 datasets. It achieves the following results on the evaluation set:
- Loss: 0.9800
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: 8
- eval_batch_size: 4
- seed: 42
- distributed_type: multi-GPU
- num_devices: 4
- gradient_accumulation_steps: 8
- total_train_batch_size: 256
- total_eval_batch_size: 16
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 2.0
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 1.0282 | 0.2122 | 100 | 1.0028 |
| 1.0122 | 0.4244 | 200 | 0.9917 |
| 0.9884 | 0.6366 | 300 | 0.9869 |
| 0.9771 | 0.8488 | 400 | 0.9841 |
| 0.9974 | 1.0610 | 500 | 0.9823 |
| 0.9934 | 1.2732 | 600 | 0.9813 |
| 0.9738 | 1.4854 | 700 | 0.9805 |
| 0.9744 | 1.6976 | 800 | 0.9801 |
| 0.9887 | 1.9098 | 900 | 0.9800 |
Framework versions
- PEFT 0.12.0
- Transformers 4.43.4
- Pytorch 2.2.1
- Datasets 2.18.0
- Tokenizers 0.19.1