Downloads · 30 days
3
15% of all-time downloads
RichardErkhov/lgaalves_-_gpt2_guanaco-dolly-platypus-4bits
lgaalves_-_gpt2_guanaco-dolly-platypus-4bits is a machine learning model from RichardErkhov. Use it for the machine learning task on the model card, and read the license before you ship it in a product.
Downloads · 30 days
3
15% of all-time downloads
All-time downloads
20
Public
Parameters
126M
127 MB on disk
Likes
0
Public
Click a slice to open those files.
.safetensors127 MB · 97%
How the weights are stored.
U884.9M · 68%
From the Hugging Face model README
Quantization made by Richard Erkhov.
gpt2_guanaco-dolly-platypus - bnb 4bits
license: mit datasets:
gpt2_guanaco-dolly-platypus is an instruction fine-tuned model based on the GPT-2 transformer architecture.
| Metric | gpt2_guanaco-dolly-platypus | GPT-2 (base) |
|---|---|---|
| Avg. | 30.25 | 29.9 |
| ARC (25-shot) | 23.55 | 21.84 |
| HellaSwag (10-shot) | 31.03 | 31.6 |
| MMLU (5-shot) | 26.4 | 25.86 |
| TruthfulQA (0-shot) | 40.02 | 40.67 |
We use state-of-the-art Language Model 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.
# Use a pipeline as a high-level helper
>>> from transformers import pipeline
>>> pipe = pipeline("text-generation", model="lgaalves/gpt2_guanaco-dolly-platypus")
>>> question = "What is a large language model?"
>>> answer = pipe(question)
>>> print(answer[0]['generated_text'])
or, you can load the model direclty using:
# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM
tokenizer = AutoTokenizer.from_pretrained("lgaalves/gpt2_open-platypus")
model = AutoModelForCausalLM.from_pretrained("lgaalves/gpt2_open-platypus")
lgaalves/gpt2_guanaco-dolly-platypus was trained using 3 datasets:
lgaalves/gpt2_guanaco-dolly-platypus was instruction fine-tuned using LoRA on 1 T4 GPU on Google Colab. It took about 1 hour to train it.
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.
Detailed results can be found here
| Metric | Value |
|---|---|
| Avg. | 25.15 |
| ARC (25-shot) | 23.55 |
| HellaSwag (10-shot) | 31.03 |
| MMLU (5-shot) | 26.4 |
| TruthfulQA (0-shot) | 40.02 |
| Winogrande (5-shot) | 50.12 |
| GSM8K (5-shot) | 0.0 |
| DROP (3-shot) | 4.96 |