|
--- |
|
license: mit |
|
datasets: |
|
- garage-bAInd/Open-Platypus |
|
language: |
|
- en |
|
pipeline_tag: text-generation |
|
--- |
|
|
|
|
|
|
|
# tinyllama-1.1b-chat-v0.3_platypus |
|
|
|
**tinyllama-1.1b-chat-v0.3_platypus** is an instruction fine-tuned model based on the tinyllama transformer architecture. |
|
|
|
|
|
### Benchmark Metrics |
|
|
|
| Metric |lgaalves/tinyllama-1.1b-chat-v0.3_platypus | tinyllama-1.1b-chat-v0.3 | |
|
|-----------------------|-------|-------| |
|
| Avg. | 37.67 | **38.74** | |
|
| ARC (25-shot) | 30.29 | **35.07** | |
|
| HellaSwag (10-shot) | 55.12 | **57.7** | |
|
| MMLU (5-shot) | **26.13** | 25.53 | |
|
| TruthfulQA (0-shot) | **39.15** | 36.67 | |
|
|
|
|
|
We use state-of-the-art [Language Model Evaluation Harness](https://github.com/EleutherAI/lm-evaluation-harness) to run the benchmark tests above, using the same version as the HuggingFace LLM Leaderboard. Please see below for detailed instructions on reproducing benchmark results. |
|
|
|
### Model Details |
|
|
|
* **Trained by**: Luiz G A Alves |
|
* **Model type:** **tinyllama-1.1b-chat-v0.3_platypus** is an auto-regressive language model based on the tinyllama transformer architecture. |
|
* **Language(s)**: English |
|
|
|
### How to use: |
|
|
|
```python |
|
# Use a pipeline as a high-level helper |
|
>>> from transformers import pipeline |
|
>>> pipe = pipeline("text-generation", model="lgaalves/tinyllama-1.1b-chat-v0.3_platypus") |
|
>>> question = "What is a large language model?" |
|
>>> answer = pipe(question) |
|
>>> print(answer[0]['generated_text']) |
|
``` |
|
|
|
or, you can load the model direclty using: |
|
|
|
```python |
|
# Load model directly |
|
from transformers import AutoTokenizer, AutoModelForCausalLM |
|
|
|
tokenizer = AutoTokenizer.from_pretrained("lgaalves/tinyllama-1.1b-chat-v0.3_platypus") |
|
model = AutoModelForCausalLM.from_pretrained("lgaalves/tinyllama-1.1b-chat-v0.3_platypus") |
|
``` |
|
|
|
### Training Dataset |
|
|
|
`lgaalves/tinyllama-1.1b-chat-v0.3_platypus` trained using STEM and logic based dataset [garage-bAInd/Open-Platypus](https://huggingface.co./datasets/garage-bAInd/Open-Platypus). |
|
|
|
### Training Procedure |
|
|
|
`lgaalves/tinyllama-1.1b-chat-v0.3_platypus` was instruction fine-tuned using LoRA on 1 V100 GPU on Google Colab. It took about 43 minutes to train it. |
|
|
|
|
|
# Intended uses, limitations & biases |
|
|
|
You can use the raw model for text generation or fine-tune it to a downstream task. The model was not extensively tested and may produce false information. It contains a lot of unfiltered content from the internet, which is far from neutral. |
|
# [Open LLM Leaderboard Evaluation Results](https://huggingface.co./spaces/HuggingFaceH4/open_llm_leaderboard) |
|
Detailed results can be found [here](https://huggingface.co./datasets/open-llm-leaderboard/details_lgaalves__tinyllama-1.1b-chat-v0.3_platypus) |
|
|
|
| Metric | Value | |
|
|-----------------------|---------------------------| |
|
| Avg. | 30.28 | |
|
| ARC (25-shot) | 30.29 | |
|
| HellaSwag (10-shot) | 55.12 | |
|
| MMLU (5-shot) | 26.13 | |
|
| TruthfulQA (0-shot) | 39.15 | |
|
| Winogrande (5-shot) | 55.8 | |
|
| GSM8K (5-shot) | 0.53 | |
|
| DROP (3-shot) | 4.94 | |
|
|
