lxe
/

Cerebras-GPT-2.7B-Alpaca-SP

Text Generation

text-generation-inference

Inference Endpoints

Model card Files Files and versions Community

lxe commited on Mar 31, 2023

Commit

ae7f22e

•

1 Parent(s): e350d3b

Update README.md

Files changed (1) hide show

README.md +5 -0

README.md CHANGED Viewed

@@ -8,6 +8,11 @@ tags:
 Cerebras-GPT-2.7B-Alpaca-SP is [cerebras/Cerebras-GPT-2.7B](https://huggingface.co/cerebras/Cerebras-GPT-2.7B) finetuned on a modified [Alpaca dataset](https://github.com/tloen/alpaca-lora/blob/main/alpaca_data_cleaned.json) with a shorter prompt structure (`Human:...\n\nAssistant:`). This contains the float16 model merged with [lxe/lora-cerebras-gpt2.7b-alpaca-shortprompt](https://huggingface.co/lxe/lora-cerebras-gpt2.7b-alpaca-shortprompt) LoRA.
 ### Basic Usage:
 ```python

 Cerebras-GPT-2.7B-Alpaca-SP is [cerebras/Cerebras-GPT-2.7B](https://huggingface.co/cerebras/Cerebras-GPT-2.7B) finetuned on a modified [Alpaca dataset](https://github.com/tloen/alpaca-lora/blob/main/alpaca_data_cleaned.json) with a shorter prompt structure (`Human:...\n\nAssistant:`). This contains the float16 model merged with [lxe/lora-cerebras-gpt2.7b-alpaca-shortprompt](https://huggingface.co/lxe/lora-cerebras-gpt2.7b-alpaca-shortprompt) LoRA.
+### Limitations:
+Although mostly coherent, the model tends to hallucinate signinficantly, often producing erroneous information.
 ### Basic Usage:
 ```python