|
--- |
|
library_name: transformers |
|
license: mit |
|
base_model: deepseek-ai/DeepSeek-R1-Distill-Qwen-1.5B |
|
tags: |
|
- generated_from_trainer |
|
- gguf |
|
- quantized |
|
- inference |
|
model-index: |
|
- name: MyModel2 |
|
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. --> |
|
|
|
# MyModel2 |
|
|
|
This model is a fine-tuned version of [deepseek-ai/DeepSeek-R1-Distill-Qwen-1.5B](https://huggingface.co/deepseek-ai/DeepSeek-R1-Distill-Qwen-1.5B) on the None dataset. |
|
It achieves the following results on the evaluation set: |
|
- Loss: 0.1089 |
|
|
|
## Model description |
|
|
|
This is a fine-tuned model available in both **SafeTensors** and **GGUF** formats. The GGUF version allows efficient inference with tools like `llama.cpp` and `ctransformers`. |
|
|
|
## Intended uses & limitations |
|
|
|
This model can be used for various natural language processing tasks. However, it may have limitations based on the dataset and fine-tuning constraints. |
|
|
|
## 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: 8 |
|
- seed: 42 |
|
- optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments |
|
- lr_scheduler_type: linear |
|
- num_epochs: 5 |
|
- mixed_precision_training: Native AMP |
|
|
|
### Training results |
|
|
|
| Training Loss | Epoch | Step | Validation Loss | |
|
|:-------------:|:------:|:----:|:---------------:| |
|
| 0.9498 | 0.2693 | 500 | 0.6119 | |
|
| 0.6245 | 0.5385 | 1000 | 0.5831 | |
|
| 0.5931 | 0.8078 | 1500 | 0.5462 | |
|
| 0.561 | 1.0770 | 2000 | 0.5148 | |
|
| 0.5312 | 1.3463 | 2500 | 0.4750 | |
|
| 0.523 | 1.6155 | 3000 | 0.4421 | |
|
| 0.5121 | 1.8848 | 3500 | 0.4096 | |
|
| 0.4059 | 2.1540 | 4000 | 0.3263 | |
|
| 0.3559 | 2.4233 | 4500 | 0.2780 | |
|
| 0.3409 | 2.6925 | 5000 | 0.2367 | |
|
| 0.3352 | 2.9618 | 5500 | 0.1973 | |
|
| 0.1918 | 3.2310 | 6000 | 0.1652 | |
|
| 0.1826 | 3.5003 | 6500 | 0.1507 | |
|
| 0.1762 | 3.7695 | 7000 | 0.1360 | |
|
| 0.168 | 4.0388 | 7500 | 0.1232 | |
|
| 0.1186 | 4.3080 | 8000 | 0.1193 | |
|
| 0.1227 | 4.5773 | 8500 | 0.1134 | |
|
| 0.1273 | 4.8465 | 9000 | 0.1089 | |
|
|
|
## Inference |
|
|
|
This model supports inference via GGUF using `llama.cpp` or `ctransformers`. |
|
|
|
### **Using `llama.cpp` (CLI)** |
|
```bash |
|
git clone https://github.com/ggerganov/llama.cpp.git |
|
cd llama.cpp |
|
make -j |
|
./main -m first.gguf -p "Hello, how are you?" |
|
``` |
|
|
|
### **Using `ctransformers` (Python)** |
|
```python |
|
from ctransformers import AutoModelForCausalLM |
|
|
|
model = AutoModelForCausalLM.from_pretrained( |
|
"your_username/your_model_repo", |
|
model_file="first.gguf", |
|
model_type="llama" |
|
) |
|
|
|
output = model("Hello, how are you?") |
|
print(output) |
|
``` |
|
|
|
## Framework versions |
|
|
|
- Transformers 4.48.2 |
|
- Pytorch 2.5.1+cu124 |
|
- Datasets 3.2.0 |
|
- Tokenizers 0.21.0 |
|
|
|
|
