predibase/drop

Description: Question answering given a passage
Original dataset: https://huggingface.co/datasets/drop
---
Try querying this adapter for free in Lora Land at https://predibase.com/lora-land!
The adapter_category is Other and the name is Question Answering (drop)
---
Sample input: Given a passage, you need to accurately identify and extract relevant spans of text that answer specific questions. Provide concise and coherent responses based on the information present in the passage.\n\n### Passage: Coming off their home win over the Browns, the Ravens flew to Heinz Field for their first road game of the year, as they played a Week 4 MNF duel with the throwback-clad Pittsburgh Steelers. In the first quarter, Baltimore trailed early as Steelers kicker Jeff Reed got a 49-yard field goal. The Ravens responded with kicker Matt Stover getting a 33-yard field goal. Baltimore gained the lead in the second quarter as Stover kicked a 20-yard field goal, while rookie quarterback Joe Flacco completed his first career touchdown pass as he hooked up with TE Daniel Wilcox from 4 yards out. In the third quarter, Pittsburgh took the lead with quarterback Ben Roethlisberger completing a 38-yard TD pass to WR Santonio Holmes, along with LB James Harrison forcing a fumble from Flacco with LB LaMarr Woodley returning the fumble 7 yards for a touchdown. In the fourth quarter, the Steelers increased their lead with Reed getting a 19-yard field goal. Afterwards, the Ravens tied the game with RB Le'Ron McClain getting a 2-yard TD run. However, despite winning the coin toss in overtime, Baltimore was unable to gain ground. In the end, Pittsburgh sealed Baltimore's fate as Reed nailed the game-winning 46-yard field goal.\n### Question: How many more field goals were made in the first half than in the second?\n### Answer:
---
Sample output: 1
---
Try using this adapter yourself!

from transformers import AutoModelForCausalLM, AutoTokenizer

model_id = "mistralai/Mistral-7B-v0.1"
peft_model_id = "predibase/drop"

model = AutoModelForCausalLM.from_pretrained(model_id)
model.load_adapter(peft_model_id)