File size: 1,281 Bytes
ed1615e 29f5049 ed1615e 192228d ed1615e 29f5049 d4604b8 29f5049 a45132c 29f5049 9b1e490 29f5049 9b1e490 29f5049 9b1e490 29f5049 9b1e490 29f5049 9b1e490 29f5049 9b1e490 29f5049 9b1e490 29f5049 |
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 |
---
library_name: transformers
license: apache-2.0
basemodel: google/gemma-7b-it
---
## Model Card for Firefly-Gemma
[gemma-7B-it-firefly](https://huggingface.co/yys/gemma-7B-it-firefly) is trained based on [gemma-7b-it](https://huggingface.co/google/gemma-7b-it) to act as a helpful and harmless AI assistant.
we trained the model on [firefly-train-1.1M](https://huggingface.co/datasets/YeungNLP/firefly-train-1.1M) dataset using LoRA.
<img src="gemma-7B-it-firefly.jpg" width="250">
## Performance
we evaluate the model on [Open LLM Leaderboard](https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard)
## Usage
The chat template of our chat models is same as Official gemma-7b-it:
```text
<bos><start_of_turn>user
Write a hello world program<end_of_turn>
<start_of_turn>model
```
You can also use the following code:
```python
from transformers import AutoTokenizer, AutoModelForCausalLM
model_name_or_path = "yys/gemma-7B-it-firefly"
tokenizer = AutoTokenizer.from_pretrained(model_name_or_path)
model = AutoModelForCausalLM.from_pretrained(model_name_or_path)
input_text = "给我写一首关于机器学习的诗歌。"
input_ids = tokenizer(input_text, return_tensors="pt")
outputs = model.generate(**input_ids)
print(tokenizer.decode(outputs[0]))
```
|